j m a t e r r e s t e c h n o l . 2 0 1 5;4(3):278–282

www.jmrt .com.br

Available online at www.sciencedirect.com

Original Article

Reduction disintegration mechanism of coldbriquettes from blast furnace dust and sludge

Leandro Rocha Lemosa,∗, Saulo Henrique Freitas Seabra da Rochab,Luiz Fernando Andrade de Castroa

a Universidade Federal de Minas Gerais (UFMG), Belo Horizonte, MG, Brazilb Hochschule Ruhr West, Mülheim an der Ruhr, Germany

a r t i c l e i n f o

Article history:

Received 4 December 2014

Accepted 6 December 2014

Available online 14 January 2015

Keywords:

Reduction disintegration index

Briquette

a b s t r a c t

It is important to understand the reduction disintegration mechanism in ferriferous burden

that is used in blast furnaces. The behavior of this burden in the granular zone of this

metallurgical reactor is important for smooth operation. The objective of this work was to

prepare cold self-reducing briquettes using blast furnace dust and sludge and binders and

compare the reduction disintegration index (RDI) of these agglomerates with conventional

ferriferous burdens such as pellets, sinter and iron ore. In the present work, 25 different

mixtures were prepared to produce briquettes in two geometries: pillow and cylindrical.

The RDI value was determined for the briquettes that passed the tumbling test.

Blast furnace © 2014 Brazilian Metallurgical, Materials and Mining Association. Published by Elsevier

Editora Ltda. All rights reserved.

carbothermic reduction of iron oxide. The overall rate of the

1. Introduction

Blast furnaces require ferriferous raw materials with adequatereducibility, high cold strength, low reduction disintegrationindex (RDI), small variations in chemical composition andappropriate particle size.

In the upper regions of the blast furnace, physical degrada-tion in the ferriferous burden occurs between 300 ◦C and 700 ◦Cbecause of the reduction of hematite to porous magnetite.

During this process, volume expansion and stress relief occurbecause of the formation and propagation of cracks [1–3]. Thisexpansion is associated with the crystallographic orientation

∗ Corresponding author.E-mail addresses: leandrorocha@demet.ufmg.br, leandrorochal@ya

http://dx.doi.org/10.1016/j.jmrt.2014.12.0022238-7854/© 2014 Brazilian Metallurgical, Materials and Mining Associa

between these minerals and porosity formation [4], althoughthe crystal lattice contracts during this transformation [5,6].

Shrinkage occurs during the reduction from hematite(trigonal–hexagonal) to magnetite (isometric–hexoctahedral),with a value of 1.93%.

Self-reducing briquettes are agglomerates that containmainly iron oxide and carbonaceous material and that requireenergy input for the reduction reactions to occur.

The Boudouard reaction has a strong effect on the overall

hoo.com.br (L.R. Lemos).

self-reduction reaction is observed to increase with increas-ing carbon content and surface area and in the presence ofcatalysts of Boudouard reaction [7–9].

tion. Published by Elsevier Editora Ltda. All rights reserved.

o l . 2 0 1 5;4(3):278–282 279

sitTt

cibti

ao

purc

awt

2

TbTw

mR

tbau

tcbfti

sc

tp

MMter

Table 1 – Mixtures prepared using the briquettingprocess.

Mixture Dust (%) Sludge(%) Binders (%)

M01 80 – BBTX: 20M02 70 – BBTX: 30M03 60 – BBTX: 40M04 – 80 BBTX: 20M05 – 70 BBTX: 30M06 – 60 BBTX: 40M07 – 90 BBTX: 10M08 – 83 (wet basis) BBTL: 17M09 80 (wet basis) – BBTL: 20M10 80 – BBTL: 20M11 76 BBTL: 24M12 35 35 BBTX: 30M13 40 40 BBTX: 20M14 45 45 BBTX: 10M15 60 30 BBTX: 10M16 60 20 BBTX: 20M17 – 80 MO: 20M18 – 76 MO: 24M19 – 80 BBTL: 20M20 – 76 BBTL: 24M21 – 71 RC: 24; W: 5M22 – 70 BBTL: 30M23 – 66 BBTL: 34M24 33 33 BBTL: 34M25 49 19 RC: 19; W: 13

Table 2 – Chemical analysis of dust, sludge and mixtureM05.

Elements (%) Materials

Dust Sludge Mixture M05

Fe 52.58 21.59 41.75K 0.13 0.19 0.30Na 0.10 0.22 0.17Zn 0.11 3.05 0.77P 0.07 0.12 0.09Mg 0.20 0.53 0.30Ca 0.77 1.3 1.86Mn 0.46 0.09 0.43Al 0.67 1.21 1.24C 32.53 49.56 19.65

strength ranged from 10 kN to 100 kN.The following mixtures were used in the piston bri-

quetting: M01, M02, M03, M04, M05, M06, M07, M12, M13,

j m a t e r r e s t e c h n

The cold agglomeration process has been used in someteel plants with up to 5% of ferriferous burden, as a source ofron and carbon for the blast furnace. However, the disadvan-age is the disintegration in the granular zone in this reactor.his is due to the destruction of the binder phase during the

emperature increase [10].Use of the appropriate binder is important to provide

ohesion between the particles of the briquette and therebyncrease the mechanical strength. Organic and inorganicinders have been used by many researchers in an attempto manufacture briquettes with an adequate RDI value for usen blast furnaces.

Organic binders decompose at high temperatures (>400 ◦C)nd degrade the briquette, reducing the viscoelasticitybserved at room temperature [11–13].

Rheology phenomena may occur during the briquettingrocess depending on the force applied to the briquette. Gran-lar materials interact by contact forces in this process andesist shearing stress to a certain level, maintaining the staticonfiguration of the particles [14,15].

This study proposes to use blast furnace dust and sludge,nd organic and inorganic binders, to produce cold briquettesith appropriate RDIs for use in blast furnaces. Two briquet-

ing processes were used: roller-press and piston briquetting.

. Methodology

he sampling and sample preparation, manufacture of theriquettes and tumbling tests were performed at Blackballsechnologies GmbH in Baesweiler, Germany. The RDI testsere performed in a laboratory at a steel company.

The following steps were performed: sample preparation,anufacture of mixtures and briquettes, curing, tumbling and

DI tests.Water (W), binder BBTL, magnesium oxide (MO), refrac-

ory cement (RC) and one organic and inorganic binder calledinder BBTX were used for the preparation of 25 mixtures on

dry basis; for details, see Table 1. Each mixture was commin-ted in a ball mill (0.35 m3; 4 h).

Normalization was used in the naming of the briquettes:he letter M indicates a mixture of substances; the letter T indi-ates pillow briquettes and the letter C indicates cylindricalriquettes; the number inside the parentheses indicates theorce/pressure applied. Chemical analyses were performed onhe dust, sludge and mixture M05, with the results presentedn Table 2.

The major crystalline phases of the blast furnace dust andludge were hematite, magnetite, graphite, wollastonite andalcite.

Two types of briquetting process were used: roller and pis-on briquetting. Then, pillow and cylindrical briquettes wererepared using these processes, as shown in Figs. 1 and 2.

The following mixtures were used in the roller briquetting:01, M02, M03, M04, M05, M06, M07, M09, M10, M11, M12,

13, M14, M15, M16, M17, M18, M22, M23, M24 and M25,

otaling 21 mixtures. A BBT-BP20 machine was used in thesexperiments. The total volume of a briquette was 3.4 cm3, theoller velocity was 1.7 rpm and the screw feeder rotational

S 0.31 1.59 0.46

velocity depended on the force applied to the briquettes. The

Fig. 1 – Pillow briquettes.

280 j m a t e r r e s t e c h n o l . 2 0 1 5;4(3):278–282

Pillow briquettes

M03

T (90

kN)

M05

T (40

kN)

M11

T (50

kN)

M12

T (60

kN)

M13

T (30

kN)

M14

T (20

kN)

M16

T (60

kN)

M23

T (20

kN)

M24

T (15

kN)

M24

T (20

kN)

RD

I-1

–3.1

5 m

m (

%)

32.5

50

40

30

20

10

0

17.5

3.1

19.822.4

36.5

46.8

1.9 2.8 1.8

Fig. 2 – Cylindrical briquettes.

M14, M15, M16, M21 and M22, totaling 14 mixtures. Onecylindrical matrix, piston and a hydraulic press (20 t capac-ity) was used to prepare the cylindrical briquettes. Eachmixture was placed inside the matrix with a mass of approx-imately 35 g and subjected to seven different pressures (10,20, 50, 100, 150, 200, and 300 MPa). Three cylindrical bri-quettes were produced with each mixture for the tumblingtests.

Non-standard tumbling tests were performed on all themanufactured cylindrical and pillow briquettes samples.For each mixture and applied force, a tumbling test wasperformed. Before performing the tumbling test, sievingof each briquette sample using a 7.5 mm sieve was per-formed.

The non-standard tumble drum used had the followingspecifications: diameter = 252 mm; width = 158 mm; two inter-nal fins of 145 mm length and 25 mm width. These tests wereperformed with a sample mass of 100 g, drum rotation veloc-ity of 25 rpm and 50 rotations. For comparison, tumbling testswere performed on four samples of sinter used in a coke blastfurnace, and a cumulative value of 90% retained on the 7.5 mmsieve was taken as the standard. The curing time used in thetumbling tests was 6 days.

The RDI tests were performed according to standardISO4696-1 (Iron Ores–Static Test for Low-Temperature Reduc-tion Disintegration) in the pillow and cylindrical briquettessamples that had cumulative values greater than 90% retainedon 7.5 mm sieve (tumbling test). The curing time for the RDItests was 5 months.

In total, 17 samples were selected for the RDI tests, andvalues higher than 90% were obtained for all these samples inthe tumbling tests: M03T (90 kN), M05T (40 kN), M11T (50 kN),M12T (60 kN), M13T (30 kN), M14T (20 kN), M16T (60 kN), M23T(20 kN), M24T (15 kN), M24T (20 kN), M04C (100 MPa), M05C(200 MPa), M07C (150 MPa), M13C (150 MPa), M15C (300 MPa),M16 C (300 MPa) and M22C (50 MPa).

X-ray diffraction analyses were performed on sample M13T(30 kN) after the RDI test – core and periphery – to investigatethe major phases present. This mixture contained hematitein its composition.

To ensure the smooth operation of blast furnaces the valueof RDI-1-3,15 mm < 30% is taken as the standard for use of bri-quettes in blast furnaces. For values of this index greater than35–40% the blast furnaces’ performance is affected [16–18].RDI-1-3,15 mm values less than 27% are acceptable for a smooth

operation of blast furnaces [19].

Fig. 3 – RDI values of pillow briquettes.

3. Results and discussion

3.1. Pillow briquettes

The selection criterion for the RDI tests among the pillow bri-quettes samples that obtained cumulative values retained ona 7.5 mm sieve >90% was random to obtain 10 samples of mix-tures of different proportions.

Fig. 3 presents the RDI results for the pillow briquettes. Intotal, 70% of the samples had RDI-3,15 mm values below 30%(the reference value used in the blast furnace). The best resultswere obtained for M11T (50 kN), M23 (20 kN), M24 (15 kN) andM24T (20 kN), which contained dust, sludge and BBTL.

The best results for the briquettes containing dust, sludgeand BBTX binder were achieved for M12T (60 kN) and M13T(30 kN). For briquettes M14T (20 kN) and M16T (60 kN), whichcontained more than 40% dust, the RDI results were greaterthan 30%.

The optimal combination of sludge and BBTX binder wasM05T (40 kN). Sample M03T (90 kN), which contained dust andBBTX binder did not exhibit adequate RDI values for use in ablast furnace. The presence of dust in the briquette composi-tion appears to be unfavorable with respect to RDI.

3.2. Cylindrical briquettes

The selection criterion for the RDI tests for the cylindrical bri-quettes that obtained cumulative values retained on a 7.5 mmsieve >90% was random to obtain seven samples of mixturesof different proportions. For the same mixture, the briquettingpressure near the maximum obtained was selected.

Fig. 4 presents the RDI results for cylindrical briquettes.In total, 71.4% of these samples exhibited RDI-3,15 mm valuesbelow 30% (the reference value).

The briquettes containing mixtures of sludge and BBTXbinder, M04C (100 MPa), M05C (200 MPa) and M07C (150 MPa),with the sludge content representing more than 70% of thecontent, exhibited RDI values below 30%. The presence of

sludge and BBTX binder in the cylindrical briquettes improvedthe RDI results.

j m a t e r r e s t e c h n o l . 2 0 1 5;4(3):278–282 281

Cylindrical briquettes

M04

C (70

kN)

M05

C (140

kN)

M07

C (105

kN)

M13

C (105

kN)

M15

C (210

kN)

M16

C (210

kN)

M22

C (35

kN)

RD

I-1

–3.1

5 m

m (

%)

12.89.5

16.0 15.2

54.4

45.4

7.3

50

60

40

30

20

10

0

Fig. 4 – RDI values of cylindrical briquettes.

vRM

r

(seltnmswh

0

0

50

100

150

Inte

nsity

(cp

s)

4.84

504Å

3.85

011Å

3.34

732Å

3.03

885Å

2.96

902Å

2.53

427Å

2.42

341Å 2.

0997

8Å2.

0291

4Å1.

9165

9Å

1.71

424Å 1.

6164

8Å

1.48

775Å

1.32

716Å

1.20

522Å

200

250-Magnetite

-Calcite

-Graphite

M

M

M

M

MM

M MM

C

C

CCC

C

G

G

G

G

20 40 60

2θ (º)80

Fig. 6 – X-ray diffraction pattern of M13T (30 kN) – core.

0

0

50

100

150

Inte

nsity

(cp

s)

4.85

158Å

3.04

053Å

3.35

474Å

2.97

119Å

2.53

453Å

2.42

366Å

2.28

116Å

2.02

963Å2.

1036

5Å

1.48

638Å

1.87

650Å

1.91

368Å

1.71

569Å

1.61

374Å

1.12

425Å

1.17

288Å

200-Magnetite

-Calcite

-Graphite

MM

M

M

M

M

M

MM

C

CC

C

C

C

C

C

G

G

G

G

G

20 40 60

2θ (º)

80

In the mixtures containing dust, sludge and BBTX binder,alues greater than 40% for dust in this mixture resulted inDI values greater than 30%, such as for M15C (300 MPa) and16C (300 MPa).

The presence of sludge and BBTL, i.e., in M22C (50 MPa)esulted in RDI values below 8%.

Fig. 5 presents secondary electron images of samples M13T30 kN) – core and periphery – after the RDI test. Fig. 5(a)hows the porous magnetite particle in the briquette periph-ry, and this morphology is similar to that reported in theiterature. Fig. 5(b) shows the porous magnetite and fracturehat occurred during the transformation of hematite to mag-etite, similar to that reported in the literature [20–24]. Theajority mineral phases identified by X-ray diffraction, as

hown in Figs. 6 and 7, indicated that all the hematite phaseas transformed into magnetite because of the absence of

ematite.

Fig. 7 – X-ray diffraction pattern of M13T (30 kN) – periphery.

HV15.00 kV

DetETD

Mag5 000 x

WD12.8 mm

Spot4.0

30 μm HV15.00 kV

DetETD

Mag5 000 x

WD14.8 mm

Spot4.0

30 μm

a b

Fig. 5 – Secondary electron image of briquette M13T (30 kN) – core (a) and periphery (b) – magnification 5000×.

n o l

r

282 j m a t e r r e s t e c h

4. Conclusions

The conclusions that were drawn from the RDI experimentsin cold self-reducing briquettes are:

• The presence of sludge and BBTX binder in the briquettesgenerated adequate RDI results (<30%);

• Dust, sludge and BBTX binder, with dust contents greaterthan 40%, in the briquettes generated RDI values greaterthan 30%. For dust contents greater than 40% the phys-ical integrity of the briquettes appears to worsen in RDItests;

• In total, 70% of the pillow briquette samples and 71.4% ofthe cylindrical briquettes samples exhibited RDI values lessthan 30%;

• The physical and chemical composition of the raw materialsof the mixture and the pressure applied to the mixture inpreparing the briquettes are important factors in attainingadequate RDI values for use in blast furnaces;

• BBTL is an adequate binder for the mixtures because of thelow RDI values;

• The mechanical strength of the briquettes is relatedto the effectiveness of the contact between the par-ticles and binders, fracture toughness of the phasespresent and homogeneous distribution of the binder in themixture.

Conflicts of interest

The authors declare no conflicts of interest.

Acknowledgments

The authors thank CAPES and DAAD for the financial supportprovided to Leandro Rocha Lemos for traveling and accommo-dation in Germany. They also thank CNPq, CAPES-PROEX andFAPEMIG. The authors thank all the laboratories that assistedwith this project, especially Blackballs Technologies GmbH inBaesweiler, Germany.

e f e r e n c e s

[1] Cores A, Muniz M, Ferreira S, Robla JI, Mochón J. Relationshipbetween sinter properties and iron granulation index.Ironmak Steelmak 2012;39:85–94.

[2] Murakami T, Kamiy Y, Kodaira T, Kasai E. Reductiondisintegration mechanism of iron ore sinter under high H2 eH2O condition at 773 K. In: 6th International congress on thescience and technology of ironmaking. Proceedings of the13th Brazilian symposium on iron ore. 2012.p. 1225–32.

[3] Loo CE, Wan KT, Howes VR. Mechanical properties of naturaland synthetic mineral phases in sinters having varyingreduction degradation indices. Ironmak Steelmak1988;15:279–85.

. 2 0 1 5;4(3):278–282

[4] Vyver WFVD, Pistorious PC, Brand SA, Reid M. Disintegrationof northern cape iron ores under reducing conditions.Ironmak Steelmak 2009;36:354–62.

[5] Paananen T [Doctoral thesis] The effect of minor oxidecomponents on reduction of iron ore agglomerates. Finland:University of Oulu; 2013.

[6] Pimenta HP, Seshadri V. Influence of Al2O3 and TiO2 onreduction degradation behaviour of sinter and hematite atlow temperatures. Ironmak Steelmak 2002;29:175–9.

[7] Yang J, Mori T, Kuwabara M. Mechanism of carbothermicreduction of hematite in hematite–carbon composite pellets.ISIJ Int 2007;47:1394–400.

[8] Moon J, Sahajwalla V. Investigation into the role of theBoudouard reaction in self-reducing iron oxide and carbonbriquettes. Metall Mater Trans B 2007;37B:215–21.

[9] Moon J, Sahajwalla V. Kinetic model for the uniformconversion of self reducing iron oxide and carbon briquettes.ISIJ Int 2003;43:1136–42.

[10] Singh M, Björkman B. Effect of processing parameters on theswelling behavior of cement-bonded briquettes. ISIJ Int2004;44:59–68.

[11] Srivastava U, Kawatra SK, Eisele TC. Study of organic andinorganic binder on strength of iron oxide pellets. MetallMater Trans B 2013;44B:1000–9.

[12] Sunde M [Doctoral thesis] Organic binder as a substitute forbentonite in ilmenite pelletization. Norway: NorwegianUniversity of Science and Technology; 2012.

[13] D’Abreu JC, Kohler HM, Takano C, Mourão MB. Self reducing:innovation awareness. In: 6th International congress on thescience and technology of ironmaking. Proceedings of the13th Brazilian symposium on iron ore. 2012. p. 2622–40.

[14] Chen DTN, Wen Q, Janmey PA, Crocker JC, Yodh AG. Rheologyof soft materials. Ann Rev Condens Matter Phys 2010;30:2–12.

[15] Wilhelm M. New methods for the rheologicalcharacterization of materials. Chem Eng Process: ProcessIntensif 2011;50:486–8.

[16] Loo CE, Bristow NJ. Properties of iron bearing materials undersimulated blast furnace indirect reduction conditions. Part 2:Reduction degradation. Ironmak Steelmak 1998;25:287–95.

[17] Jouhari AK, Ray HS. Use of iron and steel industry wastes insinter making. In: Proceeding of the international conferenceon environmental management in metallurgical industries.2000. p. 169–72.

[18] Kinnunen K, Laitinen P. Investigation of sinter plantproduction rate and RDI by neural networks. La RevueMétall-CIT 2005:361–71.

[19] Umedevi T, Brahmacharyulu A, Roy AK, Mahapatra P, PrabhuM, Ranjan M. Influence of iron ore fines feed size onmicrostructure: productivity and quality of iron ore sinter.ISIJ Int 2011;51:922–9.

[20] Rao YK, Moinpour M. Kinetics of reduction of hematite withhydrogen gas at modest temperatures. Metall Trans B1983;14B:711–23.

[21] Ünal A, Bradshaw AV. Rate process and structural changes ingaseous reduction of hematite particles to magnetite. MetallTrans B 1983;14B:743–52.

[22] Swann PR, Tighe NJ. High voltage microscopy of reduction ofhematite to magnetite. Metall Trans B 1977;8B:479–87.

[23] Bradshaw AV, Matyas AG. Structural changes and kinetics inthe gaseous reduction of hematite. Metall Trans B

1976;7B:81–7.

[24] Huang Z, YI L, Jiang T. Mechanisms of strength decrease inthe initial reduction of iron ore oxide pellets. Powder Technol2012;221:284–91.