Low Carbon Sinter Making: Use of Inert Gas to Improve Sinter RDI

Abstract

Sinter making is an intermediate step in value-added chain for producing steel. Iron ore fines (−10 mm) is agglomerated to make it suitable blast furnace feed. In sintering process, iron-ore is partially reduced, and different fluxes are added to eliminate the undesirable elements in the slag. Solid fuel mainly coke breeze is used both as energy provider and reducing agent. During the sintering process, oxidised phase (Hematite) is converted to reduced phase (Magnetite). However, during post sintering process, the reduced phase again gets oxidised as it reacts with the atmosphere during cooling. Thus, higher the coke breeze input, higher is the generation of reduced phases and in turn higher % of reduced phases will be retained in the product sinter. With increase in low grade ores, consumption of fuel and fluxes will further increase. The challenge, therefore, is to produce sinter with good RDI without increasing the solid fuel. The knowledge on this domain is very limited and thus in this work an innovative approach on injecting an inert gas during cooling of sinter was suggested. The hypothesis is that the injection of inert gas reduces the availability of oxygen therefore inhibiting the re-oxidation of reduced phase to oxidised phase in sinter. Lab scale trials followed by a plant trial in TSL India sinter plant strengthened the hypothesis as the sinter RDI improved by 8 points. This technology has the potential to significantly decrease the coke breeze requirement and thereby reduce CO₂ emissions, increased mine life, without affecting the sinter quality.

1. Introduction

A large percentage of iron ore fines generated in the steel industry are predominantly recycled via the agglomeration making route. Sintering is a generic term that is used to describe a high temperature process in which a raw materials mixture is converted into a form of agglomerate known as sinter. Iron ore sinters are suitable porous materials obtained by agglomeration processes of iron ore fine particles. These agglomerates are widely employed in blast furnaces to produce pig iron. Enough cold handling strength of sinter is required to minimise fines generation and different processes are used for this purpose.¹⁾

Sinter is produced at temperatures (~1200°C) without complete melting of the iron ore particles. Thus, the microstructure of sinter consists largely of unreacted ore particles, bonded by large amount of complex ferrite phase, as well as, partially reduced hematite or magnetite phases.^2,3,4,5,6) The type and amount of flux and fuel can thus affect the sinter quality.^4,7,8) Sinters are a product of recrystallisation and partial melting of iron ore fines and slag formation. Thus, the quality of sinter is very much dependent on its microstructural characteristics. Reduction Degradation Index (RDI) of sinter is high temperature property which affects the productivity of a blast furnace. As the burden moves down the furnace there is change in phases due to presence of reducing gases in the furnace. Secondary hematite/ Primary Hematite is converted to magnetite, resulting in volume change and generation of cracks. Thus, RDI of sinter can be largely attributed to the presence of secondary hematite that is formed due to the re-oxidation of magnetite formed during sinter.^9,10,11) During sintering process, after the burn through is achieved and sintering is complete, the magnetite in sinter gets partially converted back to hematite (secondary hematite), which is detrimental to sinter quality.¹²⁾ Thus, inhibiting re-oxidation of magnetite during cooling of sinter could be useful for improving (reducing) the RDI of sinter.¹³⁾

Previous researchers have observed that RDI is not only dependent on basicity or CaO of the sinter.^14,15) MgO also stabilises magnetite phase formation by suppression of hematite and Ca-ferrite phases.¹⁶⁾ The ore blends also play an important role.¹⁷⁾ The properties of the bonding phases (calcium ferrites and glass) are very important. The strength of the phases so as to resist the propagation of crack improves the sinter RDI.^18,19) Thus, sintering is a complex process and a large number of independent variables affect sinter quality and it is challenging to obtain optimum values of all these variables. This however leads to saturation in sinter quality due to availability of very few handles. At present the only control we have to increase magnetite content is by increasing the fuel rate, which has a significant impact on cost and environment. The challenge, therefore, is to produce sinter with sufficient high temperature strength without increasing the solid fuel. The knowledge on this domain is very limited and thus in this work emphasis is given to develop a suitable process where-in iron ore sinter have adequate high temperature strength.

2. Fundamental Investigation for Use of Inert Gas

To study the effect of inert gas purging on sinter RDI a laboratory scale pot sinter set-up was developed at Tata Steel Ltd. R&D centre. To simulate the plant conditions, the laboratory scale pot sinter facility was developed according to the similarity rules.

2.1. Experimental Method

2.1.1. Pot Sinter Setup

Raw materials were weighed in the requisite proportions and was then charged into a granulation drum. The rotation speed, inclination angle and moisture addition were kept constant for all the cases. Towards the end of the cooling process inert gas (Nitrogen) was purged with the help of a hood. A schematic of the pot sinter setup is as shown in Fig. 1

Fig. 1.

a) Pot sintering setup, b) Pot sintering setup with purging hood. (Online version in color.)

At present, only image analysis by optical microscopy is being carried out for phase identification and quantification. It is established from the analysis of phase quantification data generated by optical image analysis, that there is no effect of size fractions on phase variations.²⁰⁾ Thus the same methodology was used for characterisation studies. Representative sinter samples crushed to −3.15 + 1 mm were cold mounted in resin and polished to mirror finish. Microstructural properties were investigated in reflected light with Carl Zeiss Image analyser, Axio Imager M2m, equipped with CCD colour image sensing features. Approx. 100 photographs of each individual mounted block at 50X magnification were analysed to get an average phase percentage. SEM analysis for elemental phase composition was performed on the same samples with the JEOL SUPRA equipment. The product sinter chemical analysis was done by ICP-AES.

2.1.2. Plant Scale Trial Setup

To further strengthen the hypothesis one-month plant trial at TSL Sinter plant was also conducted. In house, inert gas purging hood was designed and the findings from lab-scale was replicated in the plant with upscaling factors.²¹⁾ The trial was planned at 3 levels of N₂ (2000 Nm³/hr, 4000 Nm³/hr, and 5000 Nm³/hr) for 10 days duration each. Samples are subjected to Chemical, RI, RDI, and TI analysis. The measurement method for RI (Reducibility Index) was IS 8167, method for RDI (Reduction Degradation Index) was JIS M8720. The tumbler index (TI) was measured as per the Indian Standard, IS 6495.

3. Methodology

3.1.1. Pot Sinter Experiments

For comparison and maintain uniformity, input raw materials (iron ore, fluxes and coke breeze) were used from the same lot. The chemistry of raw materials is shown in Table 1. Raw materials were weighed in the proportions shown in to get the desired sinter chemistry. The raw materials were mixed into the granulation drum. The total amount of moisture used was 6% of base mix and the mixing time was 15 minutes.

Table 1. Ratio and Chemistry of raw materials.

Material used	Weight%	Chemistry, wt%
		Fe(t)	Al2O3	SiO2	CaO	MgO	P	S	LOI
Iron ore	59.5	64.5	1.70	1.64	0.15	0.05	0.07	0.03	3.37
Limestone	13.9	1.26	0.25	1.38	52.90	0.55	0.07	0.04	42.06
Burnt Lime	2	0	0	0.46	74.64	1.34	NA	NA	22.54
Pyroxenite	3.7	3.9	1.0	35.0	0	37.3	NA	NA	10.13
Return fines	14.9	55.85	2.38	4.79	10.7	1.94	0.03	0.02	0

Basicity~2.5, MgO~2% and Solid fuel ~6. %; The sum of the constituent including Solid fuel is 100%

Green mix from the granulation drum was transferred to the sinter pot of 200 mm diameter and filled upto a height of 600 mm. The top of the green mix was then ignited using the movable ignition hood fuelled by L P gas while suction is applied across the bed. Ignition temperature was 1100°C while suction during ignition was maintained at 5.88 kPa. Ignition continued for 2 minutes after which the ignition hood was removed, and the suction increased to 14.7 kPa. The increased suction was maintained throughout the experiment. The experiment was considered to be complete when the wind-box thermocouple recorded the maximum temperature. This experiment was considered as the base case.

In inert gas purging tests, when the temperature of waste gas reached its maximum, a hood with nozzles connected to the inert gas line was placed on the pot and gas flow (50 lpm) was started and monitored with the help of a rotameter. The results from these experiments will now be further referred as expt with inert gas.

Three experiments were conducted for each case to check for reproducibility.

The sinter cake after cooling was stabilised by dropping it onto a steel plate. Stabilized sinter was then screened into various size fractions and tested for its micro and strength parameters (RI, RDI and TI). Chemical analysis was done using ICP-AES. Quantitative and Qualitative phase analysis of pot sinter was done by image analysis system.

3.1.2. Plant Scale Trials

Once the hypothesis was tested and proven in laboratory scale, upgradation of the process for subsequent plant scale trials were required. In first stage trial with 3 levels of N₂ was conducted. The plant selected didn’t have N₂ supply line; hence, trial was done with the temporary arrangements of N₂ tankers and vaporisers. Liquid N₂ was vaporised to a gaseous state and then charged into the sinter-making process. This trial produced the desired results with a potential to decrease solid fuel.

The raw material ratio, especially the solid fuel rate was kept constant before and during the trial period. No variation was observed in process parameters for the duration.

4. Results and Discussion

Inhibiting re-oxidation of magnetite during cooling could be useful for improving the RDI of sinter. An innovative approach on injecting an inert gas after attaining burn through is proposed. The injection of inert gas reduces the _PO₂ and therefore inhibits the re-oxidation of magnetite in the bottom part of the bed. The sinter FeO level and high magnetite: hematite ratio suggests that the temperature at the bottom of the bed may be considerably higher than 1200°C.

To evaluate the above hypothesis, pot sintering experiments and plant scale trials were conducted by purging inert gas at the end of sintering. Results are discussed in the following sections.

4.1. Pot Sintering Results

Three experiments were conducted at each level to check for reproducibility. The procedure for the pot sinter tests is the same as described in Section 2.1.1 and 3.1.1.

The procedure for stabilisation of sinter in these tests was slightly different than that mentioned in Section 3.1.1. A pot which can be split was used for the experiments. The split pot was opened, and the sinter cake was divided into three parts (Top, Middle and Bottom) along the height of the bed. Each part was separately stabilised and screened as per the procedure mentioned in Section 3.1.1. Sinter from the three portions was tested for its chemical constituents, RDI and RI. Sinter from each portion was also subjected to microstructural analysis.

4.1.1. Pot Experiment Results

There was no substantial change in TI of sinter. It varied between 64–65. This is also expected as the overall chemistry, raw material ratio and solid fuel was kept same for each experiment. The chemical analysis of sinter from the three portions of each test is reported in Table 2. There did not seem to be any definitive trend in the FeO of sinter from top to bottom in the individual tests.

Table 2. Chemical analysis of sinter from the tests.

Test Condition	Test Number	Sample from	Fe(T)	FeO	CaO	SiO2	Al2O3	MgO
Base Case	Test 1	Top	57.35	11.84	7.85	4.39	2.37	1.56
		Middle	57.81	15.74	9.51	4.98	2.40	1.68
		Bottom	56.98	13.64	9.65	5.19	2.43	1.96
	Test 2	Top	58.11	15.22	8.66	5.04	2.65	1.75
		Middle	56.83	12.09	9.12	4.86	2.49	1.64
		Bottom	56.51	13.21	9.85	4.99	2.49	2.03
	Test 3	Top	57.39	12.58	9.79	4.67	2.41	1.84
		Middle	56.68	11.14	10.35	4.72	2.41	2.22
		Bottom	57.80	11.48	9.60	4.43	2.36	1.83
Inert gas injection tests	Test 1	Top	56.33	12.87	9.68	4.90	2.55	1.69
		Middle	57.16	12.69	9.03	4.94	2.59	1.83
		Bottom	58.84	11.60	7.87	4.55	2.43	1.54
	Test 2	Top	56.21	13.35	9.22	5.16	2.64	1.94
		Middle	57.42	14.19	9.56	5.14	2.43	1.79
		Bottom	57.33	12.90	9.53	4.97	2.44	1.95
	Test 3	Top	57.29	14.06	9.21	4.83	2.47	1.98
		Middle	56.00	14.19	9.79	5.14	2.54	1.79
		Bottom	56.14	13.03	9.96	4.96	2.50	1.64

The average FeO of sinter for the base case and the inert gas injection tests are shown in Fig. 2. From the figure, it appears that the average FeO slightly decreased from top to bottom in both the base case and inert gas injection tests. However, this inference may not be accurate due to the high standard deviation associated with the average value of FeO.

Fig. 2.

Variation in FeO of sinter from top to bottom in base case and inert gas injection tests (Error bars indicate standard deviation). (Online version in color.)

The layer-wise RDI and RI of sinter from base case and inert gas tests is shown in Figs. 3 and 4 respectively. From Fig. 3, it can be observed that there is no significant change in the RDI of sinter in the top and middle portions of the pot. However, there is a significant drop in the RDI of sinter by ~3 points in the bottom portion of the sinter bed in inert gas tests compared to the base case tests. Thus proving the hypothesis that injection of inert gas at the end of sintering improves the RDI of sinter. The flow rate of inert gas was limited to 50 lpm. Higher flow rates of inert gas may improve the RDI substantially.

Fig. 3.

Variation in RDI of sinter from top to bottom in base case and inert gas injection tests (Error bars indicate standard deviation). (Online version in color.)

Fig. 4.

Variation in RI of sinter from top to bottom in base case and inert gas injection tests (Error bars indicate standard deviation). (Online version in color.)

From Fig. 4, it can be observed that the RI of sinter increased from top to bottom in the bed and that is there is no significant difference in the RI values of sinter from the base case and inert gas injection tests.

In the case of inert gas purging, it is observed that the amount of ferrite decreased and the amount of magnetite increased at the bottom of the bed. The amount of secondary hematite also decreased slightly when inert gas is injected. Secondary hematite was distinguished from primary hematite by homogenous and tabular/rhombohedral structure, unlike primary hematite which depicts micropores and massive texture. The effect of inert atmosphere on the phases in sinter at the bottom of the bed is shown in Fig. 5. This observation is in line with the results reported by other researchers^22,23) and shown in Fig. 6. According to them, tablet samples cooled in inert gas atmosphere contained a high percentage of magnetite and lower ferrite (SFCA) compared to the samples cooled in air. The variation in the amount of phases due to inert gas injection was not found to be significant in the top and middle portions of the sinter bed.

Fig. 5.

Effect of cooling atmosphere on sinter phases at the bottom of the bed.

Fig. 6.

Effect of cooling atmosphere on sinter phases as reported by Yang and Matthews.²³⁾

4.2. Plant Trial Result

Plant trial was planned in two phases to evaluate the effect of inert gas flow rate. The inert gas was injected at the rate of 2000, 4000, and 5000, Nm³/hr for a total of approx. 30 days in first phase. Trial for the second phase was aimed at studying the effect of inert gas at reduced solid fuel in the sinter. The results for the first phase is summarised in Table 3 below.

Table 3. Plant trial results.

	Inert gas Nm3/Hr	Fe(t)%	CaO%	SiO2%	FeO%	MgO%	Al2O3%	RDI	RI	TI	C rate Kg/ Ton of sinter
Base case	0	54.24	13.07	5.01	10.35	2.04	2.36	25	76	75	39.5
Inert gas	2000	54.22	13.08	5.19	10.46	2.03	2.44	21	78	77	39.8
	4000	54.41	12.90	5.11	11.01	1.88	2.48	19	78	75	39.5
	5000	53.88	13.49	5.28	10.38	2.13	2.39	17	77	75	39.3

The amount of inert gas injected has a significant effect on sinter RDI and is evident from Fig. 7. The trial was conducted continuously, with progressive increment. Thus, maintaining the reduced _PO₂ environment in the purging zone. It can be observed that there is significant improvement in RDI with an increase in inert gas volume.

Fig. 7.

Effect of inert gas flow rate on sinter RDI.

For every 0.1% increase in alumina levels in ores, there is an increase in RDI by 1 point corresponding to an increase in fuel rate by 2.4 kg/ton of sinter.^11,19,24) The effect of alumina on sinter RDI is evident from Fig. 8. It can be observed that in the case of inert gas usage, despite high alumina, the RDI was much below normal values.

Fig. 8.

Effect of Alumina on sinter RDI: Comparison with trial data.

In the above trial the Carbon rate was kept constant. Theoretically for every one point increase in RDI required ~2.4 kg/ton of sinter of solid fuel increase.^25,26) As per the trial data RDI has dropped from 25 to 17, i.e. 8 points. Thus considering even 50% decrease in requirement of fuel for RDI correction there is a scope of reduction of 10 kg/ton of sinter solid fuel, and a corresponding decrease in CO₂ emissions.

4.2.1. Microstructural Analysis of Plant Sinter

Reduction Degradation Index (RDI) of sinter is high temperature property which affects the productivity of a blast furnace and is dependent on the amount of reducing phases available in sinter. As the burden descends inside the furnace there is change in phases due to presence of reducing gases in the blast furnace. Oxidised phases (Hematite) are converted to reduced phase (magnetite), resulting in 3.34% loss in mass²⁷⁾ leading to generation of cracks, resulting in degradation and generation of fines. The fines generation needs to be minimal as it affects the upward movement of reducing gases in BF and their reaction with sinter to help extraction of Fe from sinter, thereby affecting its productivity as well. As reported by earlier researchers,^{4,18,19,28,29)} magnetite is desirable for improving sinter RDI. SFCA phase due to its acicular structure also helps in arresting crack propagation and improve both RDI and RI.^3,30,31)

Few selected samples of varying RDI were analysed for phase quantification, and in line with the above findings, it can be observed that an increase in SFCA and Magnetite phase Fig. 9, RDI is also improved. However, RDI deteriorates with an increase in hematite.

Fig. 9.

Effect of sinter phases on RDI.

4.2.2. SEM Analysis of Phases

Samples were subjected to elemental analysis of phases by SEM. The data reported is an average of 80 data points from each sample (Table 4). One sample each of higher and lower RDI was selected for the purpose. The major phases in sinter, hematite, magnetite, SFCA and Silicates were of comparative elemental values. However in case of High RDI sinter Alumino silicate (Al Si) phase was observed (Table 4). This phase is formed towards the cooling cycle and tends to have glassy structure and is thus brittle compared to other phases Due to it’s glassy texture and higher amount of alumina this phase is prone to crack generation.^19,32) The Al₂O₃ in sinter was same through out the trial period, it might be possible that low _PO₂ (due to nitrogen purging) favoured the formation of Calcium Ferrite type of phases in case of low RDI sinter.^3,33,34,35)

Table 4. SEM analysis of sinter with different RDI value.

Phases/RDI	Elemental analysis by SEM
25RDI	Al2O3%	CaO%	Fe2O3%	MgO%	SiO2%
SFCA	5.79	19.04	70.87	0.64	3.67
Magnetite	1.15	0.52	97.43	0.78	0.12
Silicate	6.52	38.58	20.47	0.10	34.33
CalciumFerrite	4.85	44.27	49.29	0.00	1.60
Alumino Silicate	32.41	0.00	10.55	0.50	56.54
17 RDI
SFCA	4.01	17.57	74.53	0.56	3.32
Magnetite	1.61	2.72	94.00	1.67	0.00
Silicate	9.15	27.12	14.85	0.00	48.88
CalciumFerrite	5.96	48.77	42.83	0.00	2.44

5. Summary

The sinter-making process is as simple as mixing iron ore fines with coke breeze and fluxes and then heating them to form an sized agglomerate. However, the reaction taking place is very complex and determines the strength of the product. Thus, controlling the reactions and getting the desired microstructure is the key. This technology controls the atmosphere towards the end of the sintering process, thereby getting higher amount of reduced phases, with a reduced coke breeze requirement. As in future, the quality of ores and other raw materials is expected to deteriorate; this technique will help in controlling the sinter process parameters as well as to maintain the desired sinter quality without any increase in solid fuel consumption.

The major observation from the above trial are:

• Sinter with lower RDI values can be generated at the same carbon rate.

• There is potential to decrease the solid fuel consumption and thereby CO₂ emissions significantly

• There is no significant impact of inert gas injection on the TI and RI of sinter

• Higher flow rate of inert gas has the potential to further reduce the RDI of sinter, provided RI is not adversely affected.

References

• 1)   S.  Harsha Nistala,  M.  Sinha,  M.  Kumar Choudhary,  G.  Bose and  S.  Sinha: Ironmaking Steelmaking, 42 (2015), 226.
• 2)   M. I.  Pownceby and  N. A. S.  Webster: 5th Annual High Temperature Processing Symposium 2013, ed. by M. A. Rhamdhani et al., Swinburne University of Technology, Hawthorn, (2013), 39.
• 3)   N. A. S.  Webster,  M. I.  Pownceby and  I. C.  Madsen: 4th Annual High Temperature Processing Symposium 2012, Swinburne University of Technology, Hawthorn, (2012), 10.
• 4)   M.  Sinha,  S. H.  Nistala,  S.  Chandra and  T. R.  Mankhand: Ironmaking Steelmaking, 44 (2017), 92.
• 5)   H.  Noda,  H.  Yanaka,  R.  Yamamoto,  H.  Kawata and  Y.  Yamaoka: Trans. Iron Steel Inst. Jpn., 25 (1985), 1103.
• 6)   T.  van den Berg,  J.  de Villiers and  R.  Cromarty: Int. J. Miner. Process., 150 (2016), 47.
• 7)   G. L.  Zhang,  S. L.  Wu,  S. G.  Chen,  B.  Su,  Z. G.  Que and  C. G.  Hou: Int. J. Miner. Metall. Mater., 21 (2014), 962.
• 8)   M. K.  Choudhary,  D.  Bhattacharjee,  P. S.  Bannerjee and  A. K.  Lahiri: ISIJ Int., 48 (2008), 1804.
• 9)   M.  Sinha and  R. V.  Ramna: METEC InSteelCon 2011, (2011), 1.
• 10)   D.  Fernández-González,  I.  Ruiz-Bustinza,  J.  Mochón,  C.  González-Gasca and  L. F.  Verdeja: Miner. Process. Extr. Metall. Rev., 38 (2017), 254.
• 11)   R. P.  Bhagat,  S. K.  Sil,  U. S.  Chattoraj,  M. C.  Goswami,  D. P.  Singh and  B. K.  Das: Tata Search, (2006), 127.
• 12)   M. M.  Hessien,  Y.  Kashiwaya,  K.  Ishii,  M. I.  Nasr and  A. A.  El-Geassy: Ironmaking Steelmaking, 35 (2008), 191.
• 13)   N.  Sakamoto,  H.  Noda,  K.  Ichikawa,  H.  Sato and  H.  Kawata: ISIJ Int., 37 (1997), 1066.
• 14)   C.  Gavat and  G.  Constantinescu: Metalurgia Bucuresti, 35 (1983), 618.
• 15)   B.  Shi,  D.  Zhu,  J.  Pan,  X.  Liu and  S.  Li: Ironmaking Steelmaking, 47 (2020), 567.
• 16)   S. C.  Panigrahy,  P.  Verstraeten and  J.  Dilewijns: Metall. Trans. B, 15 (1984), 23.
• 17)   L. H.  Hsieh: ISIJ Int., 45 (2005), 551.
• 18)   P. R.  Dawson: Ironmaking Steelmaking, 20 (1993), 137.
• 19)   M.  Sinha and  R. V.  Ramna: ISIJ Int., 49 (2009), 719.
• 20)   M.  Sinha,  V.  Tiwari,  T. K.  Ghosh and  A. S.  Reddy: ISIJ Int., 53 (2013), 1112.
• 21)   M.  Sinha,  A.  Agarwal,  S. H.  Nistala and  T. K. S.  Kumar: A Process to Produce Quality Agglomerates of Sinter, India Patent Application No. 201631003382, (2016).
• 22)   L. X.  Yang and  E.  Matthews: ISIJ Int., 37 (1997), 854.
• 23)   L. X.  Yang and  E.  Matthews: ISIJ Int., 37 (1997), 1057.
• 24)   T.  Umadevi,  A. V.  Deodar,  P. C.  Mahapatra,  M.  Prabhu and  M.  Ranjan: Steel Res. Int., 80 (2009), 686.
• 25)   C.  Murty,  A.  De,  A.  Chatterjee and  V.  Rao: Tata Search, (1994), 7.
• 26)   K.  Kinnunen and  P.  Laitinen: IFAC Proc. Vol., 37 (2004), 97.
• 27)   Z.-y.  Chen,  Y.-x.  Qu,  C.  Zeilstra,  J.  Van Der Stel,  J.  Sietsma and  Y.-x.  Yang: J. Iron Steel Res. Int., 26 (2019), 1285.
• 28)   M.  Sinha,  S. H.  Nistala,  S.  Chandra and  T. R.  Mankhand: J. Miner. Mater. Charact. Eng., 3 (2015), 180.
• 29)   M.  Sinha,  S. H.  Nistala,  S.  Chandra,  T. R.  Mankhand and  A. K.  Ghose: Ironmaking Steelmaking, 44 (2017), 100.
• 30)   N. A. S.  Webster,  J. G.  Churchill,  F.  Tufaile,  M. I.  Pownceby,  J. R.  Manuel and  J. A.  Kimpton: ISIJ Int., 56 (2016), 1715.
• 31)   N. A. S.  Webster,  M. I.  Pownceby,  R.  Pattel,  J. R.  Manuel and  J. A.  Kimpton: ISIJ Int., 59 (2019), 263.
• 32)   Z.  Ji,  Y.  Zhao,  M.  Gan,  X.  Fan,  X.  Chen and  L.  Hu: Minerals, 9 (2019), 449.
• 33)   N. A. S.  Webster,  M. I.  Pownceby,  I. C.  Madsen and  J. A.  Kimpton: ISIJ Int., 53 (2013), 774.
• 34)   M.  Valdez,  K.  Prapakorn,  A. W.  Cramb and  S.  Sridhar: Ironmaking Steelmaking, 29 (2002), 47.
• 35)   N. A. S.  Webster,  D. P.  O’Dea,  B. G.  Ellis and  M. I.  Pownceby: ISIJ Int., 57 (2017), 41.