Blast furnace and metallurgical coke’s

reactivity and its determination by thermal

gravimetric analysis
J. I. Rodero1, J. Sancho-Gorostiaga1, M. Ordiales1, D. Fernández-González1,
J. Mochón2, I. Ruiz-Bustinza2, A. Fuentes3 and L. F. Verdeja1
The aim of the present work is to adapt tests that are typically used for blast furnace cokes, such
as coke reactivity index (CRI) and coke strength after reaction, to ferroalloy production in electric
furnaces by developing easier equipment that meets with ISO 18894 standards. Moreover, a new
technique has also been developed using thermal gravimetric analysis in order to quickly,
inexpensively and reliably find the CRI parameter. As result of this work, a polynomial relationship
between coke’s reactivity and mass loss slopes of the record line obtained in the gravimetric
thermal analysis tests was found.
Keywords: Coke reactivity index, Coke strength after reaction, Ferroalloys, Electric furnace, TGA test analysis, Manganese metallurgy

Introduction temperature reducing material. Gas flow is altered by

coke’s degradability, leading to operation and coke
The optimum coke quality required by electrical consumption problems that have been well established
furnaces used in manganese industry is not the same as in blast furnace practice and that cannot be ignored in
that required in blast furnaces used in siderurgy (related electric furnace practice.
to mechanical degradability and chemical reactivity).1 Thus, the idea is to come to a simple criterion that will
However, the manganese industry is, nowadays, work- allow us to develop a quality control system for cokes
ing with a great variety of cokes, which have the same and semicokes used in ferromanganese industry. This
quality parameters [coke reactivity index (CRI) and coke development is possible due to the fact that, nowadays,
strength after reaction (CSR)] that are used in siderurgy, the quality control system used is the same as that used
though the quality parameters that should be used for in siderurgy, CRI and CSR tests. It is our aim to relate
manganese metallurgy should not be the same as that for the CRI with thermal gravimetric mass losses (from the
siderurgy. graphic slope parameter obtained by means of thermal
Ferroalloy industry (FeMn and SiMn) uses a large gravimetric analysis), in order to get a system that will
variety of cokes; so, following blast furnace practices, allow a quality control of cokes fed to manganese
reactivity and degradability of these metallurgical cokes industry electric furnaces.
will be studied. By means of these studies, it will be Six standard cokes [provided by the Spanish Coal
known whether coke properties are suitable for achiev- Institute (INCAR-CSIC)] with known reactivity have
ing their function in electric furnaces. Some authors been taken. With these cokes, it has been proved that the
question the influence of low reactivity to CO2, even procedure and equipment developed (taking into
coke degradability is downplayed,2 as an important account ISO 18894 standard) work properly. Once the
factor for coke savings and a correct reduction of min- furnace was standardised, properties (CRI and CSR) for
erals in the manufacture of ferroalloys, as postulated for some cokes used in Spanish ferroalloy industry were
the blast furnace3 and for other furnaces. It must be calculated.
accepted, therefore, that a poor quality coke, i.e. with A simple and complete method has also been devel-
high reactivity and degradability,4,5 will not provide oped to calculate the CRI by thermal gravimetric anal-
adequate coke consumption, elevating them and wasting ysis using a 1500uC thermal balance. The aim was to find
the excess CO produced, since the electric furnace’s shaft a parameter provided by thermal gravimetric analysis,
is too short, having low distance for its roles as low which might be related to each coke’s reactivity. The
slope in the thermal weight loss line record is this
parameter.
1
Dpto C. Materiales e I. Metalúrgica, Escuela de Minas de Oviedo,
Universidad de Oviedo, C/ Independencia 13, Oviedo 33004, Spain
2
Centro Nacional de Investigaciones Metalúrgicas (CENIM-CSIC), Coke reactivity
Avda Gregorio del Amo, 8, Madrid 28040, Spain
3
Director Fábrica de Boo (Cantabria), Ferroatlántica Calle del Boo, High temperature coke reactivity can be defined as the mass
Astillero 39610, Cantabria, Spain loss produced when it reacts with oxidising agents, such as
*Corresponding author, email [email protected] CO2, water vapour, O2, etc., under certain conditions.

Ñ 2015 Institute of Materials, Minerals and Mining

Published by Maney on behalf of the Institute
Received 19 December 2014; accepted 25 March 2015
618 DOI 10.1179/1743281215Y.0000000016 Ironmaking and Steelmaking 2015 VOL 42 NO 8
 Rodero et al. Blast furnace and metallurgical coke’s reactivity

When talking about coke’s reaction with CO2 in blast significant number of cokes, where they observed that
furnace, reactivity (called carboxyreactivity) process is when coke’s reactivity was great, then coke’s resistance
regulated by the Boudouard equilibrium was less in degradability test, and vice versa.

CðsÞ þ CO2 ðgÞ ! 2COðgÞ þ DH8 ¼ 163 kJ ð1Þ

Experimental work CSR/CRI by means of
ISO 18894 standard
The functions that relate temperature with DG8 (kJ) and
DH8 (kJ) are obtained by means of HSC data software6 Equipment used to determine CRI
A furnace was built for the purpose of determining
coke’s reactivity, based on the Nippon Steel Corpor-
DH8 ¼ 0:177 2 0:0074TðKÞ ð2Þ ation method and ISO 18894 standard (Fig. 2).
The furnace that has been designed for this work is a
DG8 ¼ 171:13 2 0:175TðKÞ ð3Þ resistance furnace, with a vertical tube of 80 cm height.
It consists of a ceramic cylinder into which a 60 mm
Using equation (1), it is possible to know when Bou-
douard’s reaction becomes favourable. Boudouard’s
reaction becomes favourable from 977.9 K.
Coke temperature increases progressively from its entry
into the furnace while in contact with a gas stream con-
taining CO2, which makes the balance of the equation (1)
move to the right, increasing coke consumption. The high
reactivity of coke, i.e. the greater reaction speed with
CO2 under certain furnace (pressure and temperature
conditions), increases its specific consumption; therefore,
the coke used in a metallurgical process must have low
carboxyreactivity. Saving the differences, something
similar happens in the electric furnace.

Determination of reactivity and degradability of

coke (CRI and CSR indices)
The most common tests used in the iron blast furnace
1 Relationships of %CSR to %CRI for various cokes10
industry, and specifically in Spanish mills, are the
Nippon Steel Corporation ones, the CRI and the CSR
tests.7
Coke reactivity index is expressed as per cent weight
loss in the coke sample due to the reaction with CO2,
under determined conditions according to ISO 18894
standards. Calculations are made as follows
A2B
CRI ¼ £ 100
A
where A is the original test sample weight before reac-
tion, and B is the sample weight after reaction in CO2.
Temperature is 11008C, CO2 flow is 5 L min21 and
time is 2 h, as in ISO 18894 standard.
A good quality coke has CRI values equal or lower
than 20–30%.8
Coke strength after reaction index shows coke’s
degradability in furnace operation and gives a parameter
related to gas permeability in the blast furnace, as it
depends on the fines produced. This index is similar to
the low temperature degradation index used in iron ore
degradability studies.9 It is obtained as follows
C
CSR ¼ £ 100
B
where B is the sample weight after reaction in CO2, and C is
the sample weight of 10 mm material under tumbling.
The index is defined by the weight percentage of the
fraction larger than 10 mm, in accordance with
ISO 18894. A good quality coke has CSR values larger
than 60%.8
Some researchers have found a relationship between
reactivity (CRI) and degradability (CSR).8,10–12 Menéndez
et al.10 proposed a linear equation (Fig. 1) obtained from 2 Constructed furnace section for reactivity determination

Ironmaking and Steelmaking 2015 VOL 42 NO 8 619

Rodero et al. Blast furnace and metallurgical coke’s reactivity

diameter refractory steel tube is inserted, within which Table 2 Coke mass loss by pyrolysis at 10008C for 1 h
the coke sample is located. Coke charge lies on a base of
Coke Mass loss/%
ceramic balls,13 6 mm in diameter, which serve as dif-
fuser, on a perforated refractory steel plate located on an Coke pattern A 0
alumina tube, supported by the bottom closure cap, Coke pattern B 0
which has the gas supply pipe. Coke pattern C 0
Different coke patterns [provided by the Spanish Coal Coke pattern D 0.5
Institute (INCAR-CSIC)] were tested to verify that reac- Coke pattern E 0.2
Coke pattern F 0
tivity results obtained with this furnace were correct. Coke Blast furnace coke 1 0
generally has a particle size between 20 and 25 mm. Blast furnace coke 2 0
Metallurgical coke 1 0
Equipment used to determine CSR Metallurgical coke 2 1.4
The tumble tube used for determining degradability of Metallurgical coke 3 8.7
coke was constructed exactly as it is described in ISO Petroleum coke 1 7
Petroleum coke 2 9.1
18894 standard.

Determination of reactivity and degradability of Tests were carried out twice. In the case of cokes that
standard samples had showed losses due to pyrolysis, these losses should
Reactivity tests were carried out in the furnace described be subtracted from the total loss, for calculating the
in the section on ‘Equipment used to determine CRI’. reactivity after the ISO 18894 standard furnace test.
Operation conditions were as follows: mass, gas flow Results for reactivity are presented in Table 3.
(N2 when heating and CO2 for testing), exposure time,
temperature and other factors as expressed in ISO 18894 Coke grain size influence on reactivity
standard. Owing to the fact that some of the industrially tested
Degradability tests were carried out as in ISO 18894 cokes had grain size lower than that specified in the ISO
standard (time and tumbling speed defined in ISO 18894 18894 standard, the effects of particle size on reactivity
standard). have been studied. For this purpose, blast furnace cokes
Six coke INCAR patterns were tested in order to already characterised, blast furnace coke 1 and blast
prove the correct operation of the equipment designed. furnace coke 2 have been used for CRI determination.
As the same results were obtained (Table 1), taking into The grain size ranges from w25 mm down to 10–16 mm
account tests tolerance (two units for both CRI and fraction (see Table 4).
CSR), it is possible to say that the equipment has a As we can see in Table 4, as the grain size decreases,
correct design. the reactivity increases. These results are in agreement
with those published by Oliveira.2
Determination of reactivity and degradability of
A relationship between CRI (%) and coke grain aize
various cokes (GS) (mm) can be established.
Reactivity
Once the CRI coke furnace was standardised, seven
cokes were tested: blast furnace coke 1, blast furnace
Table 3 CRI index for cokes tested
coke 2, metallurgical coke 1, metallurgical coke 2, met-
allurgical coke 3, petroleum coke 1 and petroleum coke Coke CRI/%
2. Petroleum cokes are typically used in electric furnaces
for ferroalloys production, due to their low price. Coke pattern A 20
Nevertheless, this sort of coke has low quality and small Coke pattern B 40
grain size; this is why we need the criterion that we have Coke pattern C 55
Coke pattern D 27
proposed in this paper. Blast furnace cokes have high Coke pattern E 30
quality and are used in the iron and steel industries. Coke pattern F 35
Metallurgical cokes are used in casting. Blast furnace coke 1 30
A study was conducted previously in a thermal bal- Blast furnace coke 2 31
ance, which showed that some cokes had mass losses due Metallurgical coke 1 37
to subcoking. Because of this, tested cokes and patterns Metallurgical coke 2 32
Metallurgical coke 3 50
underwent prior pyrolysis at 1000uC for a period of 1 h
Petroleum coke 1 20
to determine mass loss. Those that were badly coked, Petroleum coke 2 19
‘semicokes’ as they are called by Oliveira,2 showed losses
reaching 10% (see Table 2).
Table 4 Blast furnace cokes 1 and 2 reactivities for varying
grain sizes
Table 1 Features of coke patterns [patterns provided by
Spanish Coal Institute (INCAR-CSIC)] CRI/%

Coke Coke Coke Coke Coke Coke Size/mm Blast furnace coke 1 Blast furnace coke 2
pattern pattern pattern pattern pattern pattern
Reference A/% B/% C/% D/% E/% F/% .25 22.1 23.3
20–25 30.4 31.1
CRI (¡2) 20 40 55 27 30 35 16–20 34.1 33.0
CSR (¡2) 68 45 20 63 58 54 10–16 37.3 38.0

620 Ironmaking and Steelmaking 2015 VOL 42 NO 8

 Rodero et al. Blast furnace and metallurgical coke’s reactivity

Blast furnace coke 1 This relationship is not applicable to petroleum cokes.

This has been concluded after testing petroleum cokes:
CRI ¼ 21:0284GS þ 51:8; R 2 ¼ 0:9459; ð4Þ petroleum coke 1 and petroleum coke 2, with pyrolysis
losses at 10008C of 7.0 and 9.1 respectively.
Blast furnace coke 2 In any case, they show low reactivity values on the
order of 20%, due to the relation between physical and
CRI ¼ 20:9618GS þ 50:8; R 2 ¼ 0:9536 ð5Þ
mineralogical properties and reactivity.14,15
From these equations, it is possible to obtain the fol- The extrapolation of the CRI/CSR relationship from
lowing approximate expression, CRI þ GS ¼ 50. There carbon cokes to petroleum ones is not valid. In fact, in
will also be a relationship between sizes and the CSR, testing the two petroleum cokes, there are great differ-
as CRI and CSR are related as discussed above. ences: 40% for petroleum coke 1 and 63% for petroleum
coke 2, not expected for similar reactivity value (20).
Degradability This confirms the prior point. Petroleum cokes are pre-
Once coke reactivity testing has been carried out, tested sented in this paper as they are used in ferroalloy man-
coke is placed into the tumble tube for calculating its ufacturing because of their low price.
degradability (CSR) (see results in Table 5).

Relationship between tested CRI and CSR Experimental work: thermal balance
At the beginning of this work, it has been seen that a study of blast furnace and metallurgical
relationship between these two parameters exists, as it is coke’s reactivity
described in the literature.10 Therefore, this relationship In 2010, Oliveira showed the relationship between the
should be fulfilled for the cokes tested. Only INCAR slopes of the electric furnace mass loss line for the
coke patterns and carbon cokes should be used because Boudouard reaction, under certain test conditions,
the two petroleum cokes have different behaviours. testing different cokes, using a common thermal balance
In Fig. 3, the relations between the obtained data are and a designed macrobalance.2
represented. Oliviera studied the possibility of obtaining the car-
In this study, using the furnace designed (taking into boxyreactivities for the cokes using thermal gravimetric
account ISO 18894 standard descriptions), a very similar analysis by determining the slope of the gravimetric loss
formula to that of the other authors was found.10,11 Small in the graph line record obtained at the beginning of the
differences may be due to the number of tested cokes. test, where the slope is constant.
CRSð%Þ ¼ 21:4953CRI þ 103:42; R 2 ¼ 0:9708 ð6Þ Effect of temperature and CO2 flow on coke
reactivity
Table 5 Coke strength after reaction index obtained in It is known that temperature, gas flow, structure, par-
trommel for cokes tested
ticle size and mineral composition affect coke reaction
Coke CSR/% rate with CO2.16–23
In this work, the effect of temperature and CO2 flow
Coke pattern A 68 has been studied using a thermal balance. First of all,
Coke pattern B 45 temperature influence will be analysed. For this purpose,
Coke pattern C 20 tests have been performed by duplicate using Princeton
Coke pattern D 63
Coke pattern E 58
coke with a flowrate of 15 L h21 CO2 and varying the
Coke pattern F 54 temperatures between 1100 and 1150uC, holding it for
Blast furnace coke 1 60 50 min (Fig. 4).
Blast furnace coke 2 58
Metallurgical coke 1 50
Metallurgical coke 2 58
Metallurgical coke 3 25
Petroleum coke 1 40
Petroleum coke 2 63

4 Temperature effect on mass loss for Princeton coke with

3 Coke reactivity index and CSR index correlation 15 L h21 CO2 flow

Ironmaking and Steelmaking 2015 VOL 42 NO 8 621

Rodero et al. Blast furnace and metallurgical coke’s reactivity

Norwich and Princeton cokes have been used for the Figure 6 shows the mass loss slope (MLS) for
study of the influence of the CO2 flow value. Mass loss the Norwich and Princeton cokes with different CO2
against time will be obtained at 1100uC in order to simulate flows.
the standard test in the vertical furnace, and varying flow
rates of 2.15, 8 and 15 L h21 will be used (Fig. 5). These Test definition in thermal balance
results agree with those reported by Dlugsz et al.17 The previous studies have served to determine the
best parameters that will define the test by the new
method:
(i) carbon dioxide flowrate: 15 L h21
(ii) heating rate up to 1100uC: 6–7uC min21, higher
heating rates would cause greater difficulties in
reaching the ‘plateau’
(iii) sample weight: *200 mg
(iv) sample particle size: between 1 and 2 mm
(v) CO2 flow sample submission time: between 35
and 65 min, depending on the needed time to get
a reliable parameter of slope
(vi) total test time: 3 h approximately.

Calculation of coke reactivity

Mass losses were obtained for some standard samples
supplied by INCAR; also other mass losses were
obtained from industrial samples tested (Fig. 7a) in the
designed CRI furnace (taking into account ISO 18894
standards descriptions). Once the thermal balance is
calibrated with those standards, 11 overall, results were
plotted in a graph obtaining a wide range of reactivities/
slopes (Fig. 8). For calculating the reactivity of a new
coke, using the thermal balance, the mass loss straight
line obtained is represented in the graph and the reac-
tivity is assessed.
Tested coke’s reactivities in Fig. 7b and c are the 11
cokes referred in Table 3, with the exception of the last
two corresponding to petroleum cokes. In Table 6, the
slope and the reactivity of the 11 cokes are considered
together.
Cokes were tested at standardised parameters, and the
MLS was calculated. Figure 8 shows the relationship
between the coke’s reactivity and the slope on the mass
loss straight lines in thermal balance, according to the
method proposed in this work. A polynomial type
relationship with a high correlation coefficient was
found.
With this values, a functional relationship between
CRI (%) and slope can be established. A polynomial

a 2.15 L h21; b 8 L h21; c 15 L h21

5 Gas flow effect CO2 on mass loss line slope for Norwich 6 Mass loss slopes for Norwich and Princeton cokes with
and Princeton cokes with flows three flows

622 Ironmaking and Steelmaking 2015 VOL 42 NO 8

 Rodero et al. Blast furnace and metallurgical coke’s reactivity

7 a mass loss for Princeton, Norwich and Burton cokes with flow of 15 L h21, b mass loss for coke patterns with flow of
15 L h21 and c mass loss for industrial cokes with flow of 15 L h21

type expression was discovered to relate the MLS and The reasons that show the high interest of the thermal
coke’s reactivity balance test compared with the vertical furnace, from
the standard, are as follows:
CRI ¼ 2 2:9745MLS2 þ 25:768MLS þ 3:7677; R 2 (i) it is possible to determine the mass loss during
¼ 0:9893 R sample heating in an N2 atmosphere before
reaching 11008C, detecting if subcoking has
Results for petroleum cokes have not been included as taken place
the mass loss does not follow a straight line but rather (ii) with this equipment, only the MLS, when react-
has an ‘s’ shape. The reasons for this are due to losses ing with CO2, is taken into account in order to
arising from pyrolysis and volatiles that produce cracks estimate reactivity; it does not take into account
and increase the grain porosity. Moreover, this effect is possible losses in mass during heating due to
enhanced because the release of volatiles occurs in poor coking; this is an advantage over the ver-
an explosive manner as it was detected in some cases. tical furnace as, in this case, these weight losses,
Therefore, this method is not considered to be before the introduction of carbon dioxide, are
suitable for determining petroleum coke’s reactivity, attributed to the reaction of coke with it; there-
as received. fore, for badly coked coals, semicokes, the ISO

Ironmaking and Steelmaking 2015 VOL 42 NO 8 623

Rodero et al. Blast furnace and metallurgical coke’s reactivity

8 Relationship between TGA and %CRI mass loss slope

9 Example of mass loss curves for well coked and badly
coked cokes

Table 6 Estimated reactivities and slope in thermal balance

A mass loss occurs: B’C’ segment, at temperature lower
for tested cokes
than 1100uC. This mass loss denotes the subcoking, and
Coke CRI/% Slope the temperature at which this occurs depends on how it
has been coked (temperature defect on heating). The
Coke pattern A 20 0.71 weight loss that occurs in this testing part, when using the
Coke pattern B 40 1.87 ISO 18894 standard, should be badly attributed to coke
Coke pattern C 55 3.08
Coke pattern D 27 0.97
reactivity.
Coke pattern E 30 1.22 With the method proposed in this paper, it is
Coke pattern F 35 1.37 possible to determine the reactivity of coke (CRI)
Blast furnace coke 1 30 1.20 avoiding the possible errors that the ISO 18894 stan-
Blast furnace coke 2 31 1.27 dard presents for badly coked cokes (semicokes),
Metallurgical coke 1 37 1.47 which is a very important circumstance when produ-
Metallurgical coke 2 32 1.32
Metallurgical coke 3 50 2.52
cing ferroalloys, as the use of different cokes is a
common practice not used in the blast furnace pig iron
production, which uses normally good quality cokes.
18894 standard produces wrong measurements Once CRI is obtained from the thermal balance, the
with a higher value of reactivity. degradability (CSR) of the coke is obtained by the
In the thermal gravimetric mass loss line, only the straight related formula.
part of the same is considered because, at the beginning,
when CO2 is introduced, there is a short curved section Repeatability of thermal balance test
that is characteristic of each coke, becoming straight The test was repeated three times in the thermal balance
immediately afterwards. to estimate the test’s repeatability (Table 7).
It can be seen that the error ranges from 0.07 to
Interpretation of thermal gravimetric graphs 3.3%. The average value of all errors is 2.2%, but even
Thermal gravimetric loss weight lines follow two the highest error value, 3.3%, is less than was expected
patterns for the blast furnace and metallurgical cokes, as to be when calculating CRI following the ISO 18894
they are well or badly coked. These two behaviours are standard.
outlined in Fig. 9.

Well coked cokes

Three segments appear in the curve for well coked cokes Table 7 Slope test repeated values, difference between
(good coking curve, Fig. 9): extremes and error
(i) segment AB: heating up the sample to 1100uC Slope
takes place without loss of mass
(ii) segment BC: keeping the temperature constant at Valor 1 Valor 2 Valor 3 Difference Error/%
1100uC, CO2 begins to be introduced; in this tran-
sition segment, the mass loss does not follow any 0.71 0.71 0.69 0.02 2.8
pattern until it stabilises when reaching the point C 0.97 0.96 0.96 0.01 1.0
(iii) segment CD: the coke begins to lose mass at a 1.20 1.18 1.17 0.03 2.5
1.22 1.21 1.20 0.02 1.6
constant rate as it is shown in the graph, so the
1.27 1.24 1.23 0.04 3.1
slope in this segment can be taken for reactivity 1.32 1.3 1.28 0.04 3.0
estimation. 1.37 1.33 1.32 0.05 3.6
1.47 1.46 1.46 0.01 0.7
Badly coked cokes 1.87 1.82 1.81 0.06 3.2
For badly coked cokes (semicokes), four segments instead 2.52 2.46 2.42 0.1 4.0
of three appear in the curve (bad coking curve, Fig. 9). 3.08 2.92 2.91 0.17 5.5

624 Ironmaking and Steelmaking 2015 VOL 42 NO 8

 Rodero et al. Blast furnace and metallurgical coke’s reactivity

Conclusions References
Using coke patterns provided by the INCAR, equip- 1. A. Cores, L. F. Verdeja, S. Ferreira, I. Ruiz-Bustinza and
ment that allows the calculation of CRI and CSR used J. Mochón: ‘Iron ore sintering. Part 1: theory and practice of
sintering process’, Dyna, 2013, 80, 152–171.
in the electric furnace, i.e. blast furnace, metallurgical 2. F. Oliveira: ‘Étude des Matériauxcarbonésutiliséscommeréducteus
and petroleum cokes, has been developed based on ISO pour la production des alliages de manganésedans le four
18894 standards. électrique’, PhD thesis, École Centrale Paris, Paris, France; 2010
The obtained CRI (%) and CSR (%) relationship is 3. S. Kamalpour and W. J. Rankin: ‘The behavior of coke in submerged
arc furnace smelting of ferromanganese’, Proc. 10th Int. Ferroalloys
Cong. on ‘Transformation through technology‘, Cape Town, South
Africa, February 2004, International Committee on Ferro-Alloys
CSR ¼ 21:4953CRI þ 103:42 (ICFA), 381–390.
4. S. Olsen, M. Tangstad and T. Lindstad: ‘Production of manganese
ferroalloys’, 1st edn; 2007, Trondheim, 6 SINTEF and Tapir
A method for determining the reactivity of the blast Academic Press.
furnace and metallurgical cokes via thermal gravimetric 5. J. Sancho, L. F. Verdeja and A. Ballester: ‘Metalurgia extractive.
analysis has been developed using mass loss straight line Procesos de obtención’ 2. 1st edn, 43–45; 2000, Madrid, Sı́ntesis.
6. A. Roine: ‘Software Outokumpu HSC chemistry thermochemical
slope in standardised conditions, which has a high re- database’; version 5.11 2002, Outokumpu, Outokumpu Research Oy.
liability, which may be an alternative to the ISO 18894 7. S. Ida, T. Nishi and H. Nakama: ‘Behaviour of burden in the
test. This new test allows detecting bad produced cokes, Higashida No. 5 blast furnace’, J. Fuel Soc. Jpn, 1971, 50, 645–654.
semicokes, not detected with ISO 18894 standards. 8. M. A. Dı́ez, R. Álvarez and C. Barriocanal: ‘Coal for metallurgical
coke production: predictions of coke quality and future
It also prevents the use of the CRI index in those cases in
requirements for cokemaking’, Int. J. Coal Geol., 2002, 50,
which it should be an erroneous value because, for 389–412.
semicokes, this parameter is increased by the lack of 9. J. Mochón, A. Cores, I. Ruiz-Bustiza, L. F. Verdeja, J. I. Robla
coking (10% in some cases). and F. Garcı́a-Carcedo: ‘Iron ore sintering. Part 2: quality indices
As the particle size of the new method is 1–2 mm (in and productivity’, Dyna, 2014, 81, 168–177.
10. J. A. Menéndez, R. Álvarez and J. J. Pis: ‘Determination of
the future, it could be the size required for industrial metallurgical coke reactivity at INCAR: NSC and ECE-INCAR
facilities to reduce specific consumptions, but it must be reactivity tests’, Ironmaking Steelmaking, 1999, 26, 117–121.
proved in the ferroalloys production electric furnaces), 11. P. Arendt, F. Huhn, H. Kuhl and G. Sbierczik: ‘CRI and CSR
CRI can be determined in industrial cokes in the market an assessment of influential factor’, Cokemaking Int., 2000, 12,
62–68.
with ,20 mm size, not able for the ISO 18894 standard.
12. V. A. Lyalyuk, A. V. Sheremet, P. L. Kekuh, A. K. Otorvin,
Petroleum cokes cannot give a CRI acceptable value D. A. Tarakanov and V. L. P. Kassim: ‘Investigation of coke
as it behaves as a semicoke. Prior pyrolysis of petroleum reactivity effect on parameters of blast furnace operation’,
cokes before being tested may not improve the result as Metall. Min. Ind., 2010, 2, 317–323.
it generates a significant amount of porosity and prob- 13. L. F. Verdeja, J. P. Sancho, A. Ballester and R. González:
‘Refractory and ceramic materials’, 1st edn; 2014, Madrid, Sı́ntesis.
ably size degradation (explosions), and this alters 14. Y. Wu, S. Wu, J. Gu and J. Gao: ‘Differences in physical properties
reactivity. and CO2 gasification reactivity between coal char and petroleum
A polynomial type expression was discovered that coke’, Process. Saf. Environ., 2009, 87, 323–330.
relates the MLS to the reactivity of coke (Fig. 8) 15. J. H. Zou, Z. J. Zhou, F. C. Wang, W. Zhang, Z. H. Dai, H. F. Liu
and Z. H. Yu: ‘Modeling reaction kinetics of petroleum coke
gasification with CO2’, Chem. Eng. Process., 2007, 46, 630–636.
16. T. Hilding, K. Kazuberns, S. Gupta, V. Sahajwalla, R. Sakurovs,
CRIð%Þ ¼ 22:9745MLS2 þ 25:768MLS þ 3:7677 B. Bjökman and J. O. Wikström: ‘Effect of temperature on coke
properties and CO2 reactivity under laboratory conditions and in
an experimental blast furnace’, Proc. Conf. on ‘Iron & steel
The differences between CRI values with the new method technology’, Charlotte, NC, USA, May 2005, Iron and Steel
and the ones that can be found using the ISO 18894 stan- Society 497–505.
17. A. Dlugsz, S. Budzyn, A. Sadowski and R. Stachura: ‘The effect of
dards are below the error of the ISO 18894 method.
temperature and CO2 concentration in reaction gas coke reactivity’,
The proposed method is not suitable to determine the Arch. Metall. Mater, 2005, 50, 977–987.
reactivity of petroleum coke, as at the moment that the 18. R. Sakurovs and L. Burke: ‘Influence of gas composition on the
CO2 is introduced, mass lose does not take place evenly; reactivity of cokes’, Fuel Process. Technol., 2011, 92, 1220–1224.
thus, its reactivity cannot be determined by its slope. 19. M. Zamalloa and T. A. Utigard: ‘Characterization of industrial
coke structures’, ISIJ Int., 1995, 35, 449–457.
We obtained a criterion that allows us to determine 20. M. Grigore, R. Sakurovs, D. French and V. Sahajwalla: ‘Coke
the quality of cokes used in manganese metallurgy gasification: the influence and behavior of inherent catalytic
electric furnaces. mineral matter’, Energy Fuels, 2009, 23, 2075–2085.
21. M. Grigore, R. Sakurovs, D. French and V. Sahajwalla: ‘Mineral
reactions during coke gasification with carbon dioxide’, Int.
Acknowledgements J. Coal Geol., 2008, 75, 213–224.
22. V. Sahajwalla, M. Dubikoba and R. Khanna: ‘Reductant
The authors would like to thank the following: Ferroa- characterization and selection: implications for ferroalloys
tlántica Boo Factory for the support of this work; processing’, Proc. 10th Int. Ferroalloys Cong. on ‘Transformation
INCAR-CSIC, particularly Drs Menéndez and Barrio- through technology’, Cape Town, South Africa, February 2004,
International Committee on Ferro-Alloys (ICFA), 351–362.
canal, for the supplier of the coke patterns and their 23. K. E. Jee: ‘The effect of coal properties on carbonization behaviour
valuable help; and Professor Sancho, Ferroatlántica and strength of coke blends’, PhD thesis, University of New South
assessor, for the coordination of the whole work. Wales, New South Wales, Australia; 2012.

Ironmaking and Steelmaking 2015 VOL 42 NO 8 625