Briquetting Conditions for Australian Hematite-Goethite Iron Ore Fines

ISIJ International, Vol. 57 (2017), ISIJ International, No. 9 Vol. 57 (2017), No. 9, pp. 1517 1523 Briquetting Conditions for Australian Hematite-Goethite Iron Ore Fines Keith Richard VINING, 1) * Jasbir KHOSA 2) and Graham J. SPARROW 3) 1) CSIRO Mineral Resources, P O Box 883, Kenmore, Queensland, 4009 Australia. 2) Xstract, 50 St Georges Terrace, Perth, Western Australia, 6000 Australia. 3) CSIRO Mineral Resources, Private Bag 10, Clayton South, Victoria, 3169 Australia. (Received on February 27, 2017; accepted on May 17, 2017; J-STAGE Advance published date: August 22, 2017) The effects of the operating parameters of the briquetting machine, feed moisture content and the addition of hydrated lime as a flux on the green and fired properties of briquettes produced from Australian iron ore fines were investigated to determine the best operating conditions for producing satisfactory briquettes. The results confirmed earlier work that iron ore fines can be agglomerated by briquetting to produce a feed material suitable for a blast furnace. Feed moisture was found to be a critical operating parameter while machine operating parameters had secondary effects. The density of the green briquettes was found to depend on feed moisture content with higher density briquettes being produced with lower moisture contents. The strength of the green briquettes depended on the density of the green briquette. Feed moisture contents of 7.5 8.5 wt% resulted in a high yield of whole briquettes with densities of 3.40 3.45 g/cm 3 and green strengths of 4.0 5.5 kgf. The briquettes produced could be dried rapidly, potentially giving higher productivity values for production of briquettes compared with pellets. During briquette induration the green briquettes performed well. The basicity of the briquettes and the firing temperature had the most significant effect on the product quality. For firing temperatures of 1 300 1 350 C, and a basicity of 1.22, fired briquettes with good mechanical strength and reduction properties were obtained. Crush strengths ranged from 200 kgf to >450 kgf. KEY WORDS: agglomeration; hematite; goethite; briquette; blast furnace; iron making. 1. Introduction At present, around 460 mt/a of unprocessed hematitegoethite iron ore fines are exported from Australia as they are technically challenging to agglomerate. Development of a process to produce a pre-agglomerated calcined product would significantly increase the value of Australian iron ore exports and reduce the energy costs associated with transport. Iron ore fines are usually agglomerated to sinter or pellets to enable them to be used as feed to iron-making blast furnaces. The agglomerated material needs to have good mechanical strength and reduction properties. Briquetting is another well-established technology that can be used to agglomerate fine particles. The briquetting process involves pressing the fines into a block or briquette of suitable size and shape in a briquetting machine, either with or without a binder. Previous work in CSIRO has shown that briquetting is a prospective alternative process for agglomerating iron ore fines. Properties of products produced from the agglom- * Corresponding author: E-mail: Keith.Vining@csiro.au DOI: http://dx.doi.org/10.2355/isijinternational.isijint-2017-052 eration of the 500 μm size fraction from an Australian hematite-goethite cyclone underflow indicated that, while better pellets may be achieved with optimum grinding, briquettes with satisfactory properties as a feed to a blast furnace can be prepared from the ore without grinding. As well, the green briquettes required little drying before firing. Thus the production of briquettes instead of pellets may offer advantages in production with respect to lower energy requirements associated with not grinding the ore, removal of free and combined moisture prior to transport and increased productivity. In an extension of previous work, the effects of operating parameters of the briquetting machine, the moisture content of the feed material and the addition of hydrated lime as a flux to the feed on the green and fired properties of the briquettes produced were investigated to determine the best operating conditions for producing satisfactory briquettes. The results are presented in this paper. 2. Experimental 2.1. Samples Samples of the 500 μm size fraction from an iron ore deposit in Western Australia was received in 200 L drums 1517

for this work. The samples were air dried to 4.62 wt% moisture, homogenised and split into smaller sub-samples of approximately 15 kg for subsequent testing and analysis. Hydrated lime ( 45 μm) was used as the flux in this work. Flux additions were made to achieve basicity levels in the range 0.3 to 2.0 as calculated by the ratio of CaO to SiO 2 contents. 2.2. Preparation of Fired Briquettes In this work, briquetting of the ore was performed using a Taiyo K-102A double roll press, described previously. This machine has two 36 mm wide, 250 mm diameter, rolls with a screw-type feeder. Of the two rolls, one is fixed and one is floating and is held against the fixed roll by an oil and gas filled ram. The briquettes produced were pillow-shaped with dimensions of 13 19 28 mm and a nominal volume of 4 cc. The feed to the briquetting machines was prepared by blending the ore with the required amount of flux and mixing for two min in a twin-screw sigma mixer. Then the required amount of water for the test was added and mixing was continued for a further ten min. The green briquettes were fired in a muffle furnace at 1 310 C for 10 min. The muffle furnace was 140 mm high, 200 mm wide and 300 mm deep. The radiant heat was provided by six 10 mm diameter silicon carbide elements drawing 20 amps on a 2-phase 240 V system. The briquettes were fired in batches of up to thirty briquettes. The briquettes were dried as they were heated up to firing temperature, a time of approximately 40 min. Firing at the specified temperature and time was followed by cooling to ~200 C over one hour. Green briquettes were also fired in a CSIRO pilot scale pot-grate designed to simulate a straight-grate induration process. The pot (Fig. 1) was 280 mm in diameter, with a 500 mm green briquette bed height and a 150 mm deep hearth layer of fired briquettes to protect the grate. The temperatures monitored were the inlet gas temperature, top of the bed, middle of the bed, and bottom of the bed. The oxygen content in the heating gas to the bed was kept at 18% during the firing cycle and the pressure drop was maintained at 900 mm H 2 O throughout the entire process cycle. An initial inlet gas temperature of 760 C was chosen for drying based on the heat sensitivity of the briquettes. The gas temperature was raised from 760 C to a firing temperature of 1 340 C when the temperature at the interface between the hearth layer and the green bed (HGBI) had reached 200 C. The burner was turned off when the HGBI temperature had reached 1 260 C, and the fired briquettes were up-draught cooled. The total time for firing was estimated from the start at 760 C until the top of the bed had cooled down to 230 C. The temperature profile through the bed ranged from a top temperature of 1 320 C to a bottom temperature of 1 270 C at the end of the firing cycle. The exit gas temperature reached was 1 080 C. 2.3. Characterisation of Briquettes While there are ISO standard test methods for determining the quality of iron ore lump, pellets and sinter, no standard methods have been developed for testing iron ore briquettes. In order to gauge the quality of briquettes produced in this work, measurements were made using ISO standard test procedures for iron ore pellets as summarised in Table 1. Measurements were made on the green briquettes before firing and after firing and reduction. Compression strengths of ten green briquettes, and ten fired briquettes, were measured using an electrically-driven compression tester with a 5 kg load cell and a closure rate of 12 mm/min. The average value was taken as the crushing strength of the sample. The heat sensitivity test was designed to assess the drying behaviour of a sample of green briquettes. Five briquettes were placed in a small gas fired reactor and subjected to hot gases at a specific temperature and flow rate that simulates the drying cycle of an induration process. Each test was run for five min duration. When no spalling was observed then another sample was taken and the temperature was increased by an increment of 20 or 50 C. A maximum temperature of 750 C is used in the test. The heat sensitivity temperature (HST) is taken as the gas temperature just below that at which spalling (catastrophic degradation) of the briquettes occurs. Reduction experiments were carried out on 500 g of briquettes by heating at 900 C for 3 h in a 70% N 2 and 30% CO gas mixture under conditions described in ISO 7215 method. 2) The swelling index during reduction was determined by measuring the volume of the briquettes before and after reduction. 3) Fig. 1. Layout of the laboratory induration pot for basket briquette firing. (A = grate bars; B = packed bed of fired briquettes; C = packed bed of green briquettes). 3. Results 3.1. Characterisation of Samples Analytical and sizing data for the ore sample used in this work are given in Tables 2 and 3, respectively. Mineralogical data, determined by the point count method of ASTM Standard E562, 4) indicated that the major components of the sample were hematite (63.3 wt%) and goethite (30.3 wt%) with a small amount of magnetite (1.2 wt%). The main gangue minerals were quartz (0.9 wt%) and kaolinite (2.6 wt%). The flux used in this work was hydrated lime. Analytical and sizing data for the flux are given in Tables 2 and 3, 1518

Table 1. Summary of tests and specifications for briquettes. Material Variable Parameter Target Test method Ore sample Size distribution Wt% passing stated size As determined Dry screening Green briquettes Fired briquettes Moisture content wt% H 2O As determined Gravimetric Density g/cm 3 Material dependent Water immersion Crushing strength kg at failure (kgf) As determined ISO 4700 1) Heat sensitivity* C As determined In-house test Crushing strength kg at failure (kgf) As determined ISO 4700 1) Reducibility wt% oxygen removed >55% preferably > 60% ISO 7215 2) Swelling index Volume change during reduction <14% ISO 4698 3) Reduced strength kg at failure (kgf) >50 kg ISO 4700 1) *In the heat sensitivity test a sample of five green briquettes was subjected to hot gases at a specific temperature and flowrate that simulates the drying cycle. The Heat Sensitivity Temperature is the gas temperature just below that which spalling (catastrophic disintegration) of the briquette occurs. Table 2. Chemical analyses (wt%) for the ore and hydrated lime flux. Component Ore Hydrated lime Fe 64.2 0.20 FeO 0.36 n.d. SiO 2 2.13 1.00 Al 2O 3 1.32 0.59 P 0.057 n.d. S 0.010 n.d. CaO 0.02 68.20 respectively. MgO 0.05 0.41 Mn 0.05 n.d. TiO 2 0.06 0.02 LOI at 371 C 3.20 n.d. LOI at 900 C 3.80 28.7 n.d. = not determined. LOI = Loss On Ignition. Table 3. Particle size data* for the ore and hydrated lime flux. Size (μm) Ore Hydrated lime 1 000 100.0 100.0 500 99.6 100.0 250 76.8 100.0 125 42.4 99.9 63 9.9 95.9 45 8.6 90.7 *Cumulative wt% passing. 3.2. Briquetting Fine Ore Briquettes were prepared at different machine settings with a range of feed moisture levels and flux additions to Fig. 2. Effect of feed moisture content and rolls speed on the yield of whole briquettes for a rolls pressure of 150 kg/cm 2 and a basicity of 0.93. the ore. Previous work indicated that the cyclone underflow ore did not need grinding to produce good briquettes with the Taiyo K-102A machine when using a rolls pressure of 90 150 kg/cm 2 and a rolls speed of 4 12 rpm. With feed moisture levels between 4 and 12 wt%, there were no difficulties in feeding the ore to the machine. 3.2.1. Effect of Machine Rolls Speed and Feed Moisture Content To determine the best machine rolls speed, ore with a basicity of 0.93 and a feed moisture content of 4 12 wt% was briquetted at a roll pressure of 150 kg/cm 2. With a rolls speed of 12 rpm the resulting agglomerates were weak and poorly formed and the yield of whole briquettes was low. As the rolls speed was reduced from 12.0 rpm, stronger briquettes and an increased yield of whole briquettes were achieved until, at the lower rolls speed, the yield and quality, of the briquettes decreased (Fig. 2). The highest yields were found to occur at feed moisture contents of 7 10 wt% with rolls speeds of 4.5 10.5 rpm. During the production of the briquettes there was a moisture loss of around 10 wt%. The density of the green briquettes increased with decreasing feed moisture content and were between 3.3 and 1519

3.5 g/cm 3 for moisture contents of 7.8 and 10.1 wt%. The crushing and breakage properties of the briquettes improved with the density of the green briquettes. Selected green briquettes were fired at 1 310 C. The crushing strengths of the fired briquettes increased with decreasing feed moisture content from 210 kgf at 12.1 wt% moisture to 447 kgf at 6.4 wt% moisture. The heat sensitivity test provides information on the drying behaviour of the green briquettes. Green briquettes with a feed moisture content of 4.5 and 6.0 wt% showed no sign of spalling up to a temperature of 750 C (the maximum temperature used), whereas thermal spalling was observed in green briquettes with a feed moisture content of over 7 wt%. To determine the best moisture content of the feed, briquettes were produced at a rolls speed of 7.0 rpm and pressure of 90 kg/cm 2 with feed moisture contents between 6 10 wt% and a basicity of 0.93. The results in Fig. 3 show the maximum yield of briquettes was obtained at a feed moisture level of 7.5 8.5 wt%. Lower yields were obtained at both higher and lower moisture contents. The crushing strength of the briquettes increased with decreasing moisture content (Fig. 3). The density of the briquettes increased from 3.2 to 3.5 g/cm 3 with increasing feed moisture content. A feed moisture content of 8.0 wt% was used in subsequent tests to determine the effects of basicity levels on the quality of the briquettes. 3.2.2. Effect of Basicity and Firing Conditions To determine the effect of basicity on the properties of the green, fired and reduced briquettes, hydrated lime was added as flux at basicity levels of 0.33 1.97 to the feed at 8.0 wt% moisture and briquettes were produced with a roll pressure of 90 kg/cm 2 and a rolls speed of 6.8 rpm. The green briquettes were fired in a muffle furnace and the properties of the fired samples determined. With these briquetting conditions, green briquettes with a density of around 3.40 g/cm 3 were obtained with a yield of 84% and a crushing strength of 7 kgf. Preliminary tests indicated that briquettes with good strength were obtained with a firing time of 10 min at 1 300 C, while heating longer for up to 30 min did not significantly increase the strength (Table 4). With a 10 min firing time, increasing the firing temperature from 1 250 C to 1 350 C increased the crushing strength of the fired briquette (from 187 kgf to 478 kgf for a basicity of 1.22). Generally, lower basicity briquettes were significantly weaker than higher basicity briquettes. On the basis of these results, a minimum firing temperature of 1 300 C, with a heating time of 10 min, was chosen as the conditions for reducing the briquettes. 3.2.3. Final Fired Properties Briquettes were fired at 1 300 C for 10 min and reduced and the results of the crushing strength of the fired briquettes, reducibility, swelling index and strength after reduction are shown in Figs. 4 and 5. The reducibility index remained relatively stable across the range of basicity levels varying from 53.8% at a basicity of 0.33 to 62.2% at a basicity of 1.58. The swelling index increased from 11.1% at the lowest basicity to 14.8% in the mid-ranges, but it decreased to zero at a basicity of 1.97. The data in Fig. 4 show that the briquettes met the target levels given in Table 1 for both parameters. The crushing strength of the briquettes after firing were all over 200 kgf with a maximum value of 367 kgf obtained at a basicity of 1.22 (Fig. 5). The crushing strength after reduction showed a strong response to changes in the basicity level, increasing from 22 kgf at a basicity of 0.33 to 121 kgf at 1.97 (Fig. 5). This change in strength is considered to be related to changes in the bonding phases of the fired briquettes. The low basicity briquettes are considered to be predominantly bonded by iron oxide-iron oxide bonds, which degrade during reduction. At increased basicity levels, slag bonding becomes more significant. These bonds Table 4. Effects of firing conditions on briquette properties* for two basicity values. Firing conditions Basicity Crushing Temperature ( C) Time strength (kgf) 0.93 1 305 1 318 1 303 10 375 1 310 20 402 1 250 10 179 1 300 10 312 1 350 10 393 1.22 1 300 1 332 1 300 10 367 1 300 20 390 1 300 30 368 1 250 10 187 1 275 10 283 1 300 10 367 1 325 10 431 Fig. 3. Effect of feed moisture content on the yield of whole briquettes and their green crushing strength for a rolls speed of 7.0 rpm, a rolls pressure of 90 kg/cm 2 and a basicity of 0.93. 1 350 10 478 *Green briquettes prepared with a rolls speed of 6.8 rpm and pressure of 90 kg/cm 2 with a feed moisture content of 8.0 wt%. Flux used was hydrated lime. 1520

Fig. 4. Effect of basicity on the reducibility and swelling index of fired briquettes prepared with a rolls speed of 6.8 rpm, a rolls pressure of 90 kg/cm 2 and feed moisture content of 8.0 wt%. are more stable during reduction, accounting for the higher reduced strengths and little or no swell at a basicity of 1.97. Chemical assays for fired briquettes with different basicities are given in Table 5. The grade of the fired briquettes varied from 63.8% Fe at a basicity of 1.97 to 65.9% Fe at a basicity of 0.33, reflecting the level of flux (hydrated lime) addition. 3.3. Pot Firing Tests Briquettes produced with a feed moisture level of 8.0 wt% moisture and a basicity of 1.22 were fired in a 280 mm diameter pot using three heating profiles to determine the effects of the drying, firing and cooling stages on the properties of the fired briquettes. The different profiles were obtained by changing the heating rates and so the drying, firing and cooling times. Details of the profiles and the properties of the fired briquettes are given in Table 6. The temperature profile through the bed ranged from a top temperature of 1 320 C to a bottom temperature of 1 270 C at the end of the firing cycle with an exit gas temperature reached of 1 080 C. Profile A had the shortest heating profile of 43.4 min. This test was conducted to target both good breakage characteristics and high productivity. It had a short drying time of 3.5 min followed by a firing time of 25.4 min and a cooling time of 14.5 min. The rapid drying cycle did not Table 5. Chemical analyses (wt%) for the fired briquettes. Basicity* 0.33 0.61 0.93 1.22 1.58 1.97 Fe 65.9 65.9 65.3 64.5 64.2 63.8 FeO 0.20 0.18 0.20 0.15 0.10 0.20 SiO 2 2.41 2.44 2.44 2.46 2.40 2.33 Al 2O 3 1.56 1.55 1.54 1.55 1.50 1.51 P 0.063 0.062 0.062 0.062 0.061 0.062 Fig. 5. Effect of basicity on the crushing strength and reduced strength of fired briquettes prepared with a rolls speed of 6.8 rpm, a rolls pressure of 90 kg/cm 2 and feed moisture content of 8.0 wt%. CaO 0.81 1.49 2.28 3.01 3.80 4.65 MgO 0.07 0.08 0.07 0.09 0.08 0.08 *Basicity was adjusted with hydrated lime. Profile no. Feed moisture (wt%) Basicity Table 6. Briquette induration results for four firing profiles. Drying time Firing conditions* Firing time Cooling time Total time Fired briquette properties Crushing strength (kgf) A 8 1.22 3.5 25.4 14.5 43.4 Top 167 Middle 265 Bottom 214 B 8 1.22 10.0 26.7 21.5 58.2 Top 305 Middle 275 Bottom 232 C 8 1.22 16.0 21.7 18.5 56.2 Top 265 Middle 207 Bottom 232 D 6 0.93 5.6 17.6 22.5 49.0 Top 245 Middle 249 Bottom 185 *The drying temperature was 760 C. At the end of firing the temperature at the top of the bed was 1 320 C and 1 270 C at the bottom with the exit gas at 1 080 C. The bed was cooled to 250 C. 1521

cause spalling problems, but resulted in a weak top layer, possibly due to thermal shock. The crushing strength of the fired briquettes were poor at the top of the bed, satisfactory in the middle, and below target strength of >230 kgf at the bottom of the bed. Profile B was used to reduce thermal shock while still targeting acceptable breakage characteristics and productivity. A steady increase of 130 C/min over a period of 10 min was adopted for the drying stage. This was followed by a firing time of 26.7 min and a cooling time of 21.5 min for a total time of 58.2 min. The drying cycle did not cause any spalling problems or thermal shock. As for profile A, some deformation and degradation of briquettes at the bottom of the bed was observed with profile B. With a feed of 8.0 wt% moisture, the briquettes at the bottom of the bed appeared to become saturated, weakened and then deformed. The crushing strength results were good for the top and middle sections of the bed, and just in the target range at the bottom. (The deformed briquettes were not included in the tests for briquettes from the bottom layer.) To address the observed briquette distortion at the bottom of the bed with profile B, an additional stage was added in profile C. The first stage became 6 min of updraft drying at 300 C to strengthen the bottom of the bed followed by 10 min of down draft drying. A steady ramp up rate of 130 C/ min was adopted for the 10 min down draft drying cycle. The bed was then fired for 21.7 min until the bottom of the bed reached 1 270 C. The final stage was to cool the bed below 200 C in 18.5 min (total time 56.2 min). Profile C was designed to reduce thermal shock, and reduce saturation in the bottom layer, while still targeting good breakage characteristics, strength and productivity. The drying cycle did not cause any spalling problems or thermal shock. No deformation or degradation was observed in the briquettes at the bottom of the bed, but the middle section of the bed appeared damaged. The crushing strength results were satisfactory for the top and bottom sections, although the low crushing strength of 207 kgf reflects the damage to the middle of the bed. It is thought that this damage was caused by condensation being driven up from the bottom of the bed in the updraft drying cycle. The green briquettes used in the tests with the three firing profiles were produced with a rolls pressure of 90 kg/cm 2 and a feed moisture level of 8.0 wt%. They had an average green briquette density of 3.44 g/cm 3 and crushing strength of 6 kgf. In the initial work it was found that briquettes with a lower feed moisture content gave denser and stronger briquettes and so briquettes were prepared at a feed moisture of 6.0 wt% in an attempt to overcome the source of degradation of the briquettes during firing. Green briquettes with a density of 3.60 g/cm 3 and a crushing strength of 10 kgf were prepared using a rolls speed of 6.0 rpm and pressure of 150 kg/cm 2, basicity of 0.93 and a feed moisture of 6.5 wt%. They were fired with profile D that had a 5 min drying time, then firing for 17.6 min at 1 340 C (gas temperature) when the bottom of the bed was 1 260 C, and 22.5 min cooling time to 250 C (total time 49.0 min). No spalling of the briquettes was observed during the firing and the results in Table 6 show that good crushing strengths were obtained for the top and middle of the bed, but the bottom of the bed the value was lower, probably due to the lower temperature achieved. The observation of the full pot firings and the results in Table 6 indicate that, for deep bed firing processes, feed moisture should be maintained at a relatively low level and green density kept relatively high to prevent spalling during firing. Also from these tests it appears that the briquettes require a steady drying rate to avoid low strength due to thermal shock. A rate of 130 C/min was found to result in no significant thermal shock. For a bed approximately 500 mm thick, the optimum firing time at a gas temperature of 1 340 C was between 22 and 26 min. Allowing the bottom of the bed to reach an increased temperature or adoption of a shallow bed can be expected to improve the bottom bed strength. However, such solutions may have a negative impact on productivity, although this is yet to be confirmed. These results indicate that several operating factors can impact on the production of acceptable fired briquettes and in further work it is proposed to carry out an experimental design for the potential controlling factors. 4. Discussion This work has confirmed earlier results that briquetting can be used to agglomerate iron ore fines to produce a feed material suitable for an iron-producing blast furnace. Good quality green briquettes with a nominal volume of 4 cc were consistently produced from a laboratory-scale 250 mm rolls briquetter under specific operating conditions. Feed moisture was found to be a critical operating parameter in terms of briquetting performance and briquette quality, while rolls pressure and rolls speed had secondary effects. Feed moisture contents of 7.5 8.5 wt% resulted in a high yield of whole briquettes with densities of 3.40 3.45 g/cm 3, green strengths of 4.0 5.5 kgf and the crush strength for fired briquettes ranged from 200 kgf to over 450 kgf. The results of the heat sensitivity tests were generally excellent, often exceeding 750 C, indicating that the briquettes could be dried rapidly, potentially giving higher productivity values for production of briquettes compared with pellets. During briquette induration in a deep bed (approximately 500 mm of green briquettes), dense strong green briquettes performed well. Excessively rapid heating to firing temperature for briquettes prepared with relatively high feed moistures (8.0 wt%) gave low strengths at the top of the bed and this was attributed to thermal shock. It was found that a firing profile with a steady ramp up to firing temperature (1 340 C gas temperature) over 10 min avoided this strength degradation. Load deformation from condensation effects occurred with briquettes made at relatively high feed moistures (8.0 wt%) and mid-range densities (3.45 g/ cm 3 ), reducing productivity levels. This did not occur with 6.0 wt% moisture feed material and higher-range densities (3.60 g/cm 3 ). The basicity of the briquettes and the temperature at which they were fired had the most significant effect on briquette properties. Firing time had minimal effect. Briquettes produced at a basicity of 1.22 had good mechanical and reduction properties while those with lower basicity levels of 0.61 0.93 had low reduced strengths. These bonds are more stable during reduction, accounting for the higher reduced strengths and little or no swell at a basicity of 1.97. 1522

Fig. 6. Operating window for production of whole briquettes from the ore using the Taiyo machine at a rolls pressure of 150 kg/cm 2. From the results of this work an operating window can be determined for briquetting the cyclone underflow ore with the Taiyo briquetting machine. The operating window for production of whole briquettes is shown in Fig. 6 for a rolls pressure of 150 kg/cm 2 and selected feed moisture levels between 4 12 wt%. To the right of the curves there is a region of low feed pressure as a result of the high rolls speed and the pockets were not filled or the briquettes were weak and split readily. To the left of the curves with the slower rolls speed there is a region where the pressure on the feed was too high and clam shelling occurred as the briquettes sheared and split along the plane of the pocket contact. Blocking of the pockets also occurred. With a density below 3.10 g/cm 3, the briquettes were too weak to withstand pocket release and either remain in the pockets or split on release. The density range of 3.10 to 3.80 g/cm 3 defines the outer limits within which whole briquettes can be formed. To determine the operating window certain product and quality parameters including yield, density, crushing strength and breakage characteristics need to be considered. Once these properties are taken into consideration, a smaller operating region can be defined for the Taiyo machine used in this work. In Fig. 6, this region occurs at rolls speeds between 5.0 and 7.5 rpm and densities between 3.40 g/cm 3 and 3.60 g/cm 3. However, in view of the numerous factors that have been found to affect the quality of the briquettes produced, further work is being undertaken to complete an experimental design with all possible factors, and this will be reported when completed. 5. Conclusions Briquetting of iron ore fines has been shown to produce a feed material suitable for direct charge to an iron-making blast furnace. Feed moisture was found to be a critical operating parameter in terms of the green briquetting performance and briquette quality, while rolls pressure and rolls speed of the briquetter had secondary effects. Feed moisture contents of 7.5 8.5 wt% resulted in a high yield of whole briquettes, hence a low recycling load, with densities of 3.40 3.45 g/cm 3 and green strengths of 4.0 5.5 kgf. The green briquettes had a high degree of thermal stability allowing for fast firing cycles and the high strength implies that they could be fired at bed depths exceeding that of typical concentrate pellets. Briquettes, prepared with a basicity of 1.22, when fired at 1 300 1 350 C for 10 min, had crushing strengths that ranged from 200 kgf to over 450 kgf with a reducibility value of 60% and a swelling index of 14.8%. Acknowledgements The contributions of colleagues in CSIRO, namely John Gannon, Celeste Salter and Ross Meakins in the design and execution of the experimental work presented in this paper are gratefully acknowledged. Robe River Mining is acknowledged for the iron ore sample used in this work and for funding provided for the study. REFERENCES 1) ISO 4700: 1996, Iron Ore Pellets Determination of Crushing Strength. 2) ISO 7215: 1995, Iron Ore Pellets Determination of Relative Reducibility. 3) ISO 4698: 1994, Iron Ore Pellets Determination of Relative Free- Swelling Index. 4) ASTM E562-11: 2011, Standard Test Method for Determining Volume Fraction by Systematic Manual Point Count. 1523