Reduction of FeO in EAF Steelmaking Slag by Metallurgical Coke and Waste Plastics Blends

Transcription

1 , pp Reduction of FeO in EAF Steelmaking Slag by Metallurgical Coke and Waste Plastics Blends James Ransford DANKWAH, 1) Pramod KOSHY, 1) Narendra M. SAHA-CHAUDHURY, 1) Paul O KANE, ) Catherine SKIDMORE, ) David KNIGHTS ) and Veena SAHAJWALLA 1) 1) Centre for Sustainable Materials Research and Technology (SMaRT@UNSW), School of Materials Science and Engineering, The University of New South Wales, Sydney, NSW 05 Australia. jrdankwah@student.unsw.edu.au ) OneSteel Sydney Mill, Rooty Hill, NSW 766 Australia. (Received on September 15, 010; accepted on November 18, 010) In Australia, the use of plastics has increased tremendously over the last few decades, but less than 0% of the waste plastics are recycled. The rest is usually landfilled, which poses major environmental problems. The solution to this problem involves the development of novel environmentally-benign technologies that would utilise these waste materials. This work investigates the reduction of EAF slags (47% FeO) by blends of metallurgical coke with High-Density Polyethylene (HDPE) plastics at C. The experiments were conducted in a laboratory-scale horizontal tube furnace, and were coupled with off-gas analysis using an infrared gas analyser and a multiple gas chromatographic analyser. The results indicate that the rate of FeO reduction in slags is significantly higher when the coke/plastics blends were used compared to pure coke, with the maximum rate of reduction (Blend 4) being over twice that of coke. Moreover, the CO content in the off-gas was observed to decrease (by 75%) with increase in the polymer content of the blend. Additionally, the degree of carburisation and the removal of sulphur from the metal improved considerably when the coke was blended with plastics. The observed improvements in the rates of reduction, carburisation and desulphurisation are attributed to the reactions of hydrogen evolved from the waste plastics at these high temperatures. KEY WORDS: FeO reduction; metallurgical coke; waste plastics; carburisation; desulphurisation; hydrogen. 1. Introduction The use of plastics has increased tremendously over the last few decades, with less than 0% being recycled in Australia. 1) The rest of these waste plastics are usually landfilled, creating environmental problems. Our previous research 5) showed that waste polymers can be used to partially replace coke as the carbon source in EAF steelmaking, with significant savings in electricity and carbon usage. As carbon and hydrogen are the major constituents of waste plastics, these clearly have the potential to be used as an alternative to coke in electric arc furnace steelmaking. Coke making is one of the largest sources of greenhouse gas emissions in the steelmaking process, 6,7) with the amount of coke injected in electric arc furnaces varying between 10 0 kg/tonne of steel produced. This amounts to an annual consumption rate of thousand tonnes for a typical mini-mill. Therefore it is critical that the coke consumption should be decreased further to curtail the emission of greenhouse gases. Ongoing research by our group indicates that a significant decrease in coke consumption could be achieved by injecting supplementary carbonaceous materials such as waste polymers into the EAF steelmaking process. However, further research is required to clarify the effects of varying amounts of the polymer in the blend on the FeO reduction, particularly with regard to the role of different off-gas constituents on the reactions. This would help to improve the process efficiency and thus reduce its environmental impact. 7) The utilisation of waste plastics as a supplementary raw material with coke and coal has generated tremendous interest in recent years, culminating in the development of innovative technologies in the power industry and for blast furnaces (JFE & Nippon Steel, Japan and Bremen Steel works, Germany). 7 10) Asanuma et al. 11) investigated the use of waste plastics in coal-plastic mixtures in NKK Keihin s blast furnace and observed that plastics had higher combustion and gasification efficiencies compared with pulverized coal and thus could be effectively utilized both as a reducing agent and as a source of fuel in blast furnaces. As plastics generally have higher hydrogen content than coke/coal, these could also help to reduce the overall CO emissions. Sahajwalla et al. 1) reported that, blends of coke and waste plastics could be effectively used to enhance slag foaming in EAF steelmaking. Tadashi s 13) investigation on the use of waste plastics as a carbon source in EAF steelmaking showed that it would lower the electrical power consumption and increase the carbon levels in the steel melt. The main aim of this investigation is to determine the effect of HDPE on the: rate of reduction of FeO in EAF slag CO emissions 011 ISIJ 498

2 degree of carburisation of the metallic iron produced content of sulphur in the metallic iron produced. Our previous research 14,15) involved the use of the sessile drop method for investigating the reduction and slag foaming behaviour of slags by blends of coke and waste polymers. However, the sessile drop method is suitable for lower proportions ( 40%) of the polymer in the blend. In this paper the method used involves the preparation of an iron oxide/carbon composite (capable of handling up to 100% polymer) which contains finely ground EAF slags uniformly blended with a mixture of coke and waste plastics. The cylindrical pellets formed by compaction were allowed to undergo reduction under inert atmosphere in a custommade horizontal tube furnace at EAF steelmaking temperatures (1 550 C) Reactions Occurring During the FeO Reduction Process The reactions that are believed to occur at steelmaking temperatures ( C) are divided into the following four categories: 1) Conversion of polymers into hydrocarbons (mainly CH 4 ) polymers C n H m(g)...(1) ) Decomposition of hydrocarbons into carbon and hydrogen: m CH nc H n m(g) (s)...() (g) The hydrocarbon could also act as a sink for CO gas, producing CO and H (Eq. (3)) m C H nco nco H...(3) n m(g) (g) (g) (g) 3) Reduction of iron oxide by hydrogen, carbon and carbon monoxide: FeO H Fe H O (g) (l) (g)...(4) 4) Auxiliary reactions 16) i) Water gas shift reaction: HO H CO...(5)...(6)...(7) ii) Reaction of water vapour with C produced from the cracking of C n H m : HO C H CO (g) (s) (g) (g)...(8) iii) iv) FeO C Fe CO (s) (l) (g) FeO CO Fe CO (g) (l) (g) (g) (g) (g) Boudouard reaction: CO C CO (g) (s) (g)...(9) Carbon dissolution into molten metal: C C...(10) (s) 5) The direct reduction of FeO in slag by hydrocarbon (C n H m ) generated from a coke/polymer blend is also possible and may be represented by Eq. (11). 1 m FeO C H Fe CO H n m n n...(11) Fig. 1. The reduction of the slag by the coke alone and in the presence of the HDPE is illustrated in Figs. 1(a) and 1(b) respectively. As shown in Fig. 1, when coke alone is used the expected reductants are C and CO. In the presence of the polymer, additional reductants in the form of H, CH 4 and other hydrocarbons (C n H m ) are expected to be furnished into the reaction system, enhancing the overall reaction rate of the FeO in the slag. 1.. Calculation of FeO Conversion and Extent of Oxygen Removal from the Slag Since the off gas data was generated continuously, the overall rate of gas generation could be described by a continuous function R o (mol s 1 ), from which the extent of reduction can be estimated by integration. In that case the extent of reduction (P) after time t can be calculated from Eq. (1):...(1) where N o is the total amount of reducible oxygen in the slag. R o was estimated from the individual molar fluxes, J i, (mol cm s 1 ) as shown in Eq. (13): R A ( J J J ) A J...(13) where A is reaction area (cm ). Assuming that the gases behave ideally, the flux of any gas component in the off-gas can be represented by Eq. (14) (room temperature 0 C). J 73 i F % 1 i Ar...(14) %Ar A Therefore, Schematic illustration of the (a) iron oxide carbon reaction and (b) the enhanced reactions of FeO in the presence of hydrocarbons generated from coke/plastics blends. P 1 R N 0 o CO H O CO FeO dn ( ) FeO t R o dt o A J...(15) %CO R F %CO %H O o Ar %Ar %Ar %Ar...(16) Where: n FeO, and F Ar are the amount of FeO (mol) at time t and argon flow rate (F Ar 1 L/min), respectively. The carbon gasification reactions (Eqs. (8) and (9)) convert any H O and CO produced through Eqs. (4) and (6) into H and CO, respectively. Thus the end-product is predominantly metallic iron and synthesis gas (CO and H ). Accordingly, and as indicated by measurements from both the IR and GC analysers, the rate of reaction could be effec- t o dt FeO ISIJ

3 Table 4. Composition of EAF slag used for the investigation. Fig.. Blend compositions of the carbonaceous materials utilised in this investigation. Table 1. Composition of HDPE. Fig. 3. (a) Cylindrical pellet (slag carbonaceous material) placed in the LECO crucible prior to the experiments (b) spherical particles of reduced metal after reaction. Table 3. Table. Composition of metallurgical coke. Ash analysis of metallurgical coke used in current study. tively modelled as a function of the content of CO and CO (for the pure coke) and CO (for the pure polymer and the blends) in the off-gas without any significant errors.. Experimental.1. Sample Selection Metallurgical coke (used in EAF steelmaking) and its blends in different proportions with granulated HDPE (Fig. ) were employed in this study as the carbonaceous materials. The chemical composition (wt%) of the carbonaceous samples (coke and HDPE) and the ash analysis are given in Tables 1 3, respectively. An EAF slag with 47.1% FeO and basicity (B3) of 1.66 was provided by OneSteel Sydney Steel Mill, which was further analysed by XRF and its composition is given in Table 4... Sample Preparation The carbonaceous materials (coke and HDPE) were crushed in a jaw crusher and vibrating grinder to limit the particle size to less than 1.0 mm. They were then premixed to make a blend. The slag was subsequently mixed with the carbonaceous blends and compacted in a specially designed die to produce cylindrical pellets ( mm thick and 14 mm diameter) (Fig. 3(a)), by applying a load of 7.5 tonnes for 1 min in a hydraulic press. The mass of the composite pellet was fixed ( 4.77 g) and it was comprised of 3.83 g slag and 0.94 g carbonaceous blend so as to have a C/O molar ratio ranging from.44 to.66. This ensures that excess carbon is available in the system to allow for the reactions to reach completion. In this ratio the O refers to the oxygen content of the FeO in the slag while the C refers to the total carbon from the coke and the HDPE in the blend. The carbonaceous blend composition chart is shown in Fig.. In this work, devolatilisation prior to the introduction of the sample into the high temperature zone of the furnace was avoided to ensure that volatiles were released into the high temperature reaction zone. This was to enable our study to establish the critical role of volatiles on iron oxide reduction. Preliminary trials showed that crucibles (supplied by LECO Australia) were the most suitable for the current investigations since they were not attacked extensively by the molten slag as was observed when aluminosilicate crucibles were used. The LECO crucibles were primarily composed of zirconia (51 wt%), silica (39%) and alumina (3.7%). The experimental procedure (Fig. 4) involved three parts: the reactions in a custom-made horizontal resistance heated furnace, visual observation using a CCD camera and the off-gas analysis using a gas chromatographic analyser (SRI 8610C Chromatograph Multiple Gas #3 GC configuration equipped with a thermal conductivity conductor (TCD)) and a continuous infrared gas analyser, to monitor offgases produced by the reduction reaction. The results (gas analyses and visual imaging) were recorded in a data-logging computer. The sample assembly was inserted in the furnace, which was purged continuously with argon (1 L/min) of 99.99% purity to remove any contaminants 011 ISIJ 500

4 Fig. 4. Schematic of the horizontal tube furnace and IR gas analyser system. Fig. 6. Comparison of the volumes of CO generated from the reactions of the slag with HDPE, Met coke, and blends of coke with HDPE at C. Fig. 7. Comparison of the volumes of CO generated from the reactions of the slag with HDPE, Met coke, and blends of coke with HDPE at C. Fig. 5. Stages of the reduction process of FeO in the slag by Blend at C. and residual oxygen in the furnace. After the furnace had attained the desired hot zone temperature (1 550 C), the sample was pushed into the reaction hot zone and the reactions were monitored for 30 min. This time was selected since initial trials showed that the reactions ceased and thus there were no further changes in gas composition or degree of reduction beyond this time. The reacted carbonaceous material/slag samples were quenched by rapidly withdrawing the tray from the hot reaction zone into the cold zone of the furnace. Particles of reduced iron metal, which were clearly visible to the naked eye (Fig. 3(b)), were removed magnetically and analysed. 3. Results and Discussions 3.1. Stages of the Reduction Process Figure 5 shows the various stages of the reduction process as captured by the CCD camera for Blend along with the calculated percent reduction (%f). The image in the reactor immediately after inserting the crucible into the hot zone is shown as time t 0. As shown in Fig. 5, melting did not commence immediately. A gas halo was formed around the pellet probably due to initial devolatilisation and gasification of the carbonaceous material at the external surface of the pellet. Melting commenced after about 60 s and was complete after a period of 300 s for this blend. However the melting time generally ranged from s depending upon the amount of plastics blended with the coke. The melting time decreased with increase in the amount of plastics blended with the coke. Some reduction occurred in the solid state, reaching about 13.4% after complete melting of the slag. This period was followed by vigorous foaming that lasted for about 40 s ( s depending on the type of blend). Rapid reduction occurred within the foaming zone, rising from 13.4% after the melting process to 30.6% in 60 s. After 540 s, when foaming had died down completely, reduction of up to 98.9% was achieved. Similar behaviour was observed for the other blends. 3.. Gas Generation Rates (IR Analyser) The contents of CO and CO in the off-gas were measured continuously by an infrared (IR) gas analyser and the amounts of CO and CO removed from the slag were calculated using the off-gas data for each blend. The results are shown in Figs. 6 and 7 for CO and CO, respectively. The rate of evolution of CO increased slowly for the first 400 s for coke followed by a sharp rise, attaining a maximum value after about 570 s. CO generation for the pure polymer became significant after about 600 s, attaining its maximum value after about 740 s. On comparing the volume of gas generated for 100% coke with the blend, it can be observed that, although the amount of fixed carbon decreased, the maximum amount of CO gas generated (vol%) increased with the amount of HDPE blended with the coke ISIJ

5 Table 5. Concentration of species from the C gasification (by H O) reaction at C using Factsage 6.0. Table 6. Concentration of species from the reduction of FeO by CH 4 at C using Factsage 6.0. (Fig. 6). A possible explanation to the observed increase in the CO produced could be the highly endothermic reaction (carbon gasification by water vapour, Eq. 8), which occurs above C 17) and is therefore spontaneous at our experimental temperature of C. FeO H Fe H O (g) (l) (g)...(4) HO C H CO (g) (s) (g) (g)...(8) DG T (J/mol) 18) The free energy change can be expressed as a function of the partial pressure (activity) of the species and the temperature (Eq. (17)). pco ph ΔG ΔG RT ln...(17) a pho C At C, Factsage 6.0 software 19) predicts the following equilibrium data for reactants and products based on the reaction between a mol of solid C and a mol of gaseous H O (Table 5). Since for each blend the C/O ratio was greater than.0, the reaction products would include excess amounts of solid carbon at unit activity. 0) The partial pressures (atm) of each gas component i, p i, at a total pressure of 1 atm could be defined as: p (% i ) / ) i...(18) Accordingly, the free energy change at C could be calculated as 44.0 J. Ueki et al. 1) showed in their work that the reduction of H O to H and CO through Eq. (8) occurred by between to 6% in the temperature range to 1 00 C. Extra CO and H generation arising from Eq. (8) is expected to be higher in this investigation, since it occurs at a higher temperature (1 550 C). At this temperature the exothermic water gas shift reaction (Eq. (7), a reaction which would have resulted in a decrease in the content of CO) is not favourable. HO H CO...(7) (g) (g) (g) DG T (J/mol) 18) It is apparent from Fig. 7 that all the blends showed lower CO evolution compared to coke and that the reduction in CO evolution increased as the amount of HDPE blended with the coke increased. The relatively lower values recorded for CO compared to CO may be an indication of direct reduction of FeO by C or a dominant Boudouard reaction (Eq. (9)) and the carbon gasification reaction (Eq. (8)), respectively. Another possible reaction is the direct reduction of FeO by CH 4, which was the predominant hydrocarbon detected in the off-gas at C. CH FeO Fe H CO 4...(19) ΔG T [J/mol] pco ( ph ) a ΔG ΔG RT ln pch a...(0) At C, Factsage ) predicts the following data for reactants and products based on the reaction between 1 mol of gaseous CH 4 and 1 mol of FeO (Table 6), from which the free energy change could be calculated as J. These results from thermodynamic calculations using FactSage ) along with thermochemical data of Kubaschewski et al. 18) show that the free energy change for Eq. (19) is negative at our experimental temperature of C, and thus this reaction is feasible under our experimental conditions. A gas chromatographic analyser equipped with a thermal conductivity detector (TCD) was used to identify the offgas constituents produced from the reaction of the slag sample with each of the carbonaceous reductants. The chromatograms obtained for pure coke, HDPE and Blend 3 are shown in Fig. 8. From this figure, it is clear that for the pure polymer and the blends, the content of H in the off gas is significantly higher as seen from the dominant peak. However, the chromatogram for the pure coke shows significantly low contents of H in the off gas (Fig. 8(b), cf. Table ). Identification of the gases by the GC was done after 1 min of reduction. For complete elution of the gases of interest (up to the H O peak) about thirteen minutes was required. The concentration of FeO in the slag is expected to decrease as the reaction progressed since the existing oxygen potential is not favourable for the reduction of the other oxides in the slag. The variation of the concentration of FeO with time for each carbonaceous reductant is shown in Fig. 9. Each of the plots in Fig. 9 has 3 regions: an initial slow reduction section followed by a linear rapid reduction and finally a slow degradation region characterised by local deficiency of FeO. The pure polymer showed the largest duration of the initial slow region followed by the pure coke; the length of the initial slow region appeared to decrease with an increase in the level of plastics blended with the coke. After the complete reduction of FeO from the slag, the process did not cease; instead, the reduction of the other oxides, notably MnO and SiO commenced as their concentrations (wt%) in the slag increased. Safarian et al. ) observed a slow reduction of MnO in an initial rapid FeO reduction stage that was followed by a second rapid MnO stage. They attributed the onset of the rapid second reduction stage to the metallothermic reduction of MnO by the Fe that was produced in the initial rapid stage. Fe (MnO) Mn (FeO) (l) 18) Fe 4 FeO...(1) 011 ISIJ 50

6 Fig. 8. Gas chromatogram obtained after 60 s of reaction of EAF slag with a reductant consisting of (a) Coke (b) Blend 3, and (c) HDPE. Table 7. Time for complete reduction of FeO from the slag by each carbonaceous reductant at C. Fig. 9. Concentration of FeO in the slag as function of time for each carbonaceous reductant at C. However, the reduction behaviour of these oxides is beyond the scope of this investigation, and the progress of the reactions was followed up to the time where maximum reduction of FeO was achieved (when n o 0.05 mol) for each carbonaceous reductant. This time for complete reduction was observed to differ for each carbonaceous reductant and followed the same trend as the length of the initial slow region, as shown in Table 7. From Table 7 the time required for complete reduction of the FeO from the slag decreased with increase in the amount of HDPE blended with the coke. The relatively lower times of reduction recorded for the blends compared to pure coke may be attributed to the presence of a hydrogen environment provided by the waste plastics. Higher hydrogen contents promote fluidity at low to moderately high temperatures and gas formation at high temperatures. Hydrogen gas, apart from being a faster reducing agent than both carbon monoxide and solid carbon, 3,4) also enhances the rate of reduction of iron oxides by carbon monoxide 4) when it is added to a reduction system containing the latter. It is obvious from Eqs. (8) and (9) that the CO and H O produced from the reduction of iron oxide (Eqs. (3) and (5)) gasify any available carbon at high temperatures. As stated before, the water gas shift reaction (Eq. (7)) is only important at temperatures below C as it is exothermic. Accordingly, a series of cyclic reactions 5,6) is set up as the carbon monoxide and hydrogen (products of the carbon gasification) in turn partially react with the iron oxide and further reactions continue. These cyclic reactions persist until all the iron oxide has been reduced to metallic iron, provided there is enough carbon in the system 6) as is the case in our system. It is thus clear that the presence of hydrogen and some fixed carbon in a carbonaceous material are both essential to initiate and maintain these cyclic reactions to obtain a high rate of reduction. This essential requirement is met when metallurgical coke ( 71% fixed carbon) is blended with waste plastics (HDPE: 14 wt% H ). This clearly explains why the blends generally perform better than both pure coke and the pure polymer. The graph of the rate of reduction (mol s 1 ) of FeO in the slag by each carbonaceous reductant is plotted as a function of time in Fig. 10, using Eq. (16). The rate of reaction of FeO (Table 8) increased with the proportion of HDPE blended with the coke, with the blends recording much higher values than for pure coke or pure polymer. For 100% metallurgical coke, a value of mol s 1 was calculated as the maximum rate, while a lower value of mol s 1 was calculated for the ISIJ

7 polymer. The corresponding rates for the blends were mol s 1, mol s 1, mol s 1 and mol s 1 for blends 1,, 3 and 4, respectively. Although these rates are higher, the trend of increasing rate with HDPE addition to the coke is similar to the results obtained by Sahajwalla et al. 7) and Rahman 8) who used a different approach (sessile drop) to study the effect of HDPE additions on carbon/slag interfacial reactions. Fig. 10. Table 8. Rate of reduction of FeO in slag as a function of time at C for each carbonaceous reductant. Maximum rates of reduction for different carbonaceous blends Fraction Reacted as a Function of Time The equations controlling the various mechanisms of the reaction for a first order chemical reaction control process (complete uniform internal reduction) and mass transfer (bulk mass transfer) are given by the equations ln(1 f) kt 9,30) and f kt, 30) respectively. The change in fraction reacted, f was plotted as a function of time for pure metallurgical coke, HDPE, and the blends. The result is illustrated in Fig. 11(a). Each of the graphs has at least two linear sections, indicating a change in mechanism or reaction order as the reaction proceeds. Whereas the pure coke followed the first order rate law ln(1 f ) kt (Fig. 1) in each of its linear sections, the pure polymer and the blends followed this law only in the first 40 s and then shifted to the rate law f kt, afterwards, (Fig. 11), suggesting a shift from chemical control and a gradual decrease in the influence of the Boudouard reaction as the coke was blended with the polymer. Further investigation is underway to fully elucidate the kinetics of the reactions Nature of Metal Produced The high ash content (18.3 wt%) of the metallurgical coke used affected metal production when the coke was used as a reducing agent without blending with plastics (Fig. 13). From Fig. 13 it is clear that metal separation from slag increases with the level of waste plastics blended with the coke, an observation that may be attributed to the detrimental effects of ash on the reduction process and the clearing of a layer of ash from the reaction surface by the plastics. The clean, smooth surface of the metals produced by the HDPE is explained by the non-wetting nature of the metal/ slag layer in the absence of ash; the presence of ash would have resulted in increased adhesion to the slag as they both contain oxides Content of Carbon and Sulphur in the Produced Metal The contents of carbon and sulphur in the produced iron metal were determined using a LECO carbon/sulphur analyser; the results are shown in Figs. 14 and 15 for carburisation and sulphur removal respectively. Carburisation is lowest for the pure plastics (0.009 wt%), followed by metallurgical coke (0.650 wt%). The very low level of carburisation observed for the pure plastics compared to coke Fig. 11. Change in fraction reacted with time for a) met coke, b) HDPE, c) Blend 1 and d) Blend ISIJ 504

8 Fig. 1. The reduction of FeO in slag with (a) met coke, (b) met coke (first 180 s) at C replotted as complete uniform internal reduction controlled. Fig. 13. Nature of metal produced from reduction of slag by (a) 100% met coke (b) Blend 3 and (c) 100% HDPE at C. Fig. 14. Carburisation of the metal formed after reduction for each carbonaceous reductant at C. Fig. 15. Content of sulphur in the metal after reduction for each carbonaceous reductant at C. may be due to the absence of fixed carbon in the polymer. It increases consistently with the amount of HDPE blended with the coke till Blend 3 and then shows a decrease. The carburisation process for each of the blends is higher than either coke or HDPE alone, indicating the existence of a favourable synergistic effect between coke and HDPE towards carburisation of metallic iron. The level of sulphur in the metal showed a reverse trend to that seen for the carbon pickup, with the highest level of sulphur (0.177 wt%) found in the metal produced using pure metallurgical coke. The level of sulphur (0.17 wt%) found in the metal produced using pure HDPE was lower than that of coke but was still significant. Significantly lower values of sulphur (0.056 wt%, 0.05 wt%, 0.04 wt% and wt%) were observed for the blends 1,, 3 and 4 respectively; this again indicates the existence of a favourable synergistic effect between coke and HDPE towards the desulphurisation of ISIJ

9 Fig. 16. Relative amounts of CO generated by reactions of the slag with each carbonaceous reductant (with CO emissions from coke as the base) at C. metallic iron. The improved performance of the blends with respect to carburisation and sulphur removal may be explained as follows: For pure metallurgical coke (no hydrogen present), the metal is carburised mainly by the reaction 31,3) : CO CO C (in Fe)...() (g) In a CO H CH 4 environment (due to thermal/catalytic cracking and gasification of the plastics) the produced metal is carburised via three parallel reactions, which consists of Eq. () and the following: 31,3) H CO H O C(inFe)...(3) CH C (in Fe) H 4(g) (g) (g) (g) (g)...(4) Fruehan 30) calculated an apparent rate constant for Eq. (3) that was over five times higher than that for Eq. (). Kaspersma and Shay 33) determined the rate constant for carburization by Eq. (3) to be 16 times greater than that by Equation () and times that by Eq. (4). Karabelchtchikova 34) calculated a rate constant for Eq. (3) that was 30 and 100 times faster than Eqs. () and (4), respectively. Therefore the rate of carburisation is expected to be significantly higher for the blends than for coke alone. The metallic iron produced was spherical in each case, indicating that it was formed from the molten state. The experimental temperature of C is only slightly above the melting point of C, indicating that further alloying elements may be present in the iron obtained. However, as stated earlier, the time taken for complete melting of pellets decreased as the proportion of the polymer in the blend increased. Further investigation is underway to fully determine the effect, if any, of polymer addition on melting behaviour as well as analyse the metallic iron produced for the presence and contents of any other alloying elements. (g) 3.6. Environmental Considerations (CO Evolution) The relative amounts of CO evolved after complete reduction of the slag by each carbonaceous reductant were calculated and the results are plotted in Fig. 16. Calculations were based on the total amount of CO generated from the start of the reduction process to the point where the reduction was judged to have reached completion (indicated by the time where the amount of oxygen removed from the slag became numerically equal to the total removable oxygen corresponding to 47.1% FeO). It is clear from Fig. 16 that the highest amount of CO was produced from the reactions involving the pure coke. These values correspond to CO generation from the reduction reaction only; values do not take into account the CO evolved during the production of the metallurgical coke from coal or the CO generated to produce the electrical energy used to power the reduction process. The observed significant decrease in CO emissions with HDPE addition agrees with the observation by Matsuda et al. 0) that it is possible to utilise waste plastics to produce metallic iron without generating CO. 4. Conclusions The major findings of this investigation are: (1) Blends of waste plastics (HDPE) with metallurgical coke could be used to partly replace the conventional metallurgical coke used in EAF steelmaking as a reductant. () Significant improvements in rates and percent reduction were observed when the metallurgical coke was partly blended with HDPE. Maximum rate of reaction increased from (pure coke) to mol s 1 (Blend 4). (3) The degree of carburisation and the removal of sulphur from the metal improved considerably when the coke was blended with the plastics. Whereas the content of carbon in the metal improved from 0.65% for pure coke to 4.5% for Blend 3 the sulphur content in the metal dropped from 0.177% (pure coke) to 0.04% (Blend 3). (4) The CO content in the off-gas decreased by up to over 75%, with increase in the amount of plastics blended with the coke. (5) H gas (from polymer decomposition) is responsible for the enhanced iron oxide reduction compared to coke alone. Acknowledgements We acknowledge the financial support received from OneSteel for this project; one of the authors is grateful to the Ghana Education Trust Fund (GETFUND) for their financial support. REFERENCES 1) (005), 1. ) V. Sahajwalla, L. Hong and N. Saha-Chaudhury: Iron Steel Technol., 3 (005), No., 54. 3) V. Sahajwalla, M. Zaharia, A. Anthony, J. Lee, S. Darma, R. Khanna, N. Saha-Chaudhury, D. Knights, P. O Kane and E. Pretorius: Proc. of AISTech 007, USA, (007), CD.-I 4) R. Khanna, M. Rahman, R. Leow and V. Sahajwalla: Metall. Trans. B, 38B (007), ) M. Zaharia, V. Sahajwalla, R. Khanna, P. Koshy and P. O Kane: ISIJ Int., 49 (009), ) R. Khanna, F. McCarthy, H. Sun, N. Simento and V. Sahajwalla: Metall. Trans. B, 36B (005), ) V. Sahajwalla, M. Rahman, R. Khanna, N. Saha-Chaudhury, P. O Kane, C. Skidmore and D. Knights: Proc. of AISTech 008, USA, (008), CD.-I 8) J. K. Yeo: J. Korean Inst. Resour. Recycl., 6 (1997),. 9) J. Janz and W. Weiss: Proc. 3rd European Iron Making Cong., CRM-. VDEh, Gent, Belgium, (1996), ) Y. Iino: NKK Tech. Rev., 88 (003), ISIJ 506

10 11) M. Asanuma, T. Ariyama, M. Sato, R. Murai, T. Nonaka, I. Okochi, H. Tsukiji and K. Nemoto: ISIJ Int., 40 (000), 44. 1) V. Sahajwalla, L. Hong and N. Saha-Chaudhury: Iron Steel Technol., (006), ) T. Hattori: Electric Furnace Steel, 76 (005), 7. 14) M. Rahman, V. Sahajwalla, R. Khanna, N. Saha-Chaudhury, D. Knights and P. O Kane: Proc. AISTech 006, USA, (006), ) V. Sahajwalla, M. Rahman, R. Khanna, J. Dicker, D. Knights, P. O Kane and C. Skidmore: Proc. AISTech 007, USA, (007), ) D. Gosh, A. K. Roy and A. Gosh: ISIJ Int., 6 (1986), ) A. K. Biswas: Principles of Blast Furnace Ironmaking, Cootha Publishing House, Brisbane, Australia, (1981), ) O. Kubaschewski, C. B. Alcock and P. J. Spencer: Materials Thermochemistry, 6th Ed., Pergamon Press, Oxford, (1993), ) C. W. Bruce, A. D. Pelton, W. T. Thompson, G. Eriksson, K. Hack, P. Chartrand, S. Decterov, J. Melancon and S. Petersen: FactSage 6.0 GTT Technologies, Aachen, Germay, (009). 0) T. Matsuda, M. Hasegawa, A. Ikemura, K. Wakimoto and M. Iwase: ISIJ Int., 48 (008), ) Y. Ueki, R. Mii, K. Ohno, T. Maeda, K. Nishioka and M. Shimizu: ISIJ Int., 48 (008), ) J. Safarian, L. Kolbeinsen, M. Tangstad and G. Tranell: Metall. Mater. Trans. B, 40B (009), 99. 3) I. Sohn and R.J. Fruehan: Metall. Mater. Trans. B, 36B (005), ) H. Ono-Nakazato, T. Yonezawa and T. Usui: ISIJ Int., 43 (003), ) E. Donskoi, D. L. S. McElwain and L. J. Wibberly: Metall. Mater. Trans. B, 34B (003), 55. 6) J. Y. Shi, E. Donskoi, D. L. S. McElwain and L. J. Wibberly: Ironmaking Steelmaking, 35 (008), 3. 7) V. Sahajwalla, M. Rahman, R. Khanna, N. Saha-Chaudhury, P. O Kane, C. Skidmore and D. Knights: Steel Res. Int., 80 (009), ) M. Rahman: PhD. Thesis, UNSW-Sydney, Australia, (010). 9) E. T. Turkdogan and J. V. Vinters: Metall. Trans., 3 (197), ) E. T. Turkdogan: Physical Chemistry of High Temperature Technology, Academic Press, New York, NY, (1980), ) R. J. Fruehan: Metall. Trans., 4 (1973), 13. 3) R. J. Fruehan: Metall. Trans., 4 (1973), ) J. H. Kaspersma and R. H. Shay: J. Heat Treat., 4 (1980), 1. 34) O. Karabelchtchikova: PhD. Thesis, Worcester Polytechnic Institute, New England, USA, (1997) ISIJ