ISIJ International, Vol. 36 (1996). No. 8, pp.

991-999

Ettect of Firing Temperature and Porosity on Thermal Oonductivity

and Diftusivity of lron Ore Pellets

N. S. and VandanaRAO
SUNDARMURTI
Department of Metallurgical Engineering. Faculty of Technology and Engineering. M. S. University of Baroda, Baroda, India
390 OOI .

(Received on January 8. l996, accepted in final form on Alpril 12. 1996)

The thermal diffusivity and conductivity of iron ore pellet were determined by measuring temperature at
surface and centre of pellet during heating and applying heat balance. The effect of firing temperature of
pellets and its porosity on thermal conductivity were determined. The mineralogical changes on firing were
investigated. The results were analysed by applying model equation and on the basis of basic sintering
parameters to determine effective thermal conductivity at zero porosity with radial heat transfer.
KEYWORDS:
thermal diffusivity/conductivity measurements; thermal properties of iron ore pellet; pellet
properties; mineralogical changes on heating of pellet.

factors like porosity or firing temperature are to be

1. Introduction
investigated.
The knowledge of thermal properties of iron ore In the present investigation an attempt has been made
agglornerate at higher temperatures are essential to to measure the temperature of pellets at surface and
analyse the heat transfer in iron making processes. centre and apply principles of heat balance to analyse
Wynnyckyj and McCurdyl) examining the induration the results. The effect of bulk density and porosity of
process observed that a three fold increase in ther- pellet obtained by firing at different temperatures were
mal conductivity of pellet resulted in a ten fold increase in investigated. An attempt has been madeto explain the
their propensity to develop shrinkage cracks. The role results on the basis of existing models and phase changes
of thermal properties and heat transfer on the reaction on heating.
kinetics of magnetite coalchar composite pellets were
emphasisedby Seaton2) and co-workers. With these facts 2. Experimental
in mind an attempt was undertaken to measurethermal 2.1. Materials
properties of iron ore pellets.
Enormousliteratures The lumps of iron ore procured from Rajhara mines
are available to measurethermal
of India were crushed and ground to fine size. The size
properties of materials. The methods used by an
analysis of the powder is indicated in Table 1. Table 2
investigator was directed by some specific needs and
reasons and varied from simple steady state conduction3) Table 1. Size analysis of iron ore powder.
method to sophisticated laser flash method.4) Modified
laser flash method was reported by Ohta5) and sub- Mesh Microns c/o Present
sequently used by Akiyama6~8) to measurethermal dif-
fusivities of dense and porous iron ore agglomerates. - 100 + 170 - 152 + 80 l .90
These non-contact methods are accurate but limited to 170 + 200
- 200 300 - 88+76 6.00

flat surface geometries and axial heat flow. Abzalov et - + - 76+65 51.50
- 250 + 300 - 65+54 4.50
al.9) also determined the thermal conductivities of small
- 300 + pan - 54 36. IO
cylinders of compressedmagnetite concentrates over a
Total lO0.00
range of temperatures and heating rates.
Measurementsof thermal diffusivity and emissivity of
fired spherical hematite pellets and dried pellets using
Table 2. Chemical analysis of iron ore.
transient state conduction equation for radial heat fiow
were reported by Watt and his co-workers.l0,11) They Constituent Percentage
measuredthe temperatures at half radius and at centre
Total iron 64.0
of the pellet by inserting thermocouples to evaluate Si02 3.3
thermal diffusivity. The insertion of two thermocouples CaO l.l
in a small pellet is not desirable specially wheneffect of

991 C 1996 ISIJ

 ISIJ International, Vol. 36 (1 996). No. 8
atures of 303, 1
393, 1
473, 1
488, 1 1
523 and 548K1
fN
o
E\
for I
hr in a silicon carbide tube furnace and for higher
a),
Semp!9 N,~m' : IRONORE temperatures in platinum woundtube furnace.
u QO
c,)
(~,,

tJr,
r(1 O) r\
,,) Nl~t
c;, ~t,
(~d
o
(~~,
~'
co
2.3. Thermal Property Studies
(~J
The method used is to one suggested by Watt
similar
c't al.10,1 l)
and different analysis. The
wlth modification
o) S,Lmple N~ole : FE-ORE9SO
co
c~l methodis selected for Its suitability to spherical geometry
Lf) and radiai heat transfer. Watt measuredthe temperature

~;
t:~
co ~ a)
r\,

r\,
rh'

o cr)
ui)
Cr)
a:,
at half radius and at the centre of the pe]let and applied
'o '~'
t\ rr) ,~~ cr) transient conduction
equation to evaluate thermal
o r) r~l*
c\l
c~, ~$
(~~
c~ o diffusivlty. of a thermocouple at centre and
Positioning
~$ n r\,

at surface of pellet is more reliabie and easier than to

N
o)
(o place at any radical distance for its accurate positioning.
N SeLmpl Name FE OREi2~~ It
was reportedl
1) that
a ten percent error in positioning
Lf)
of thermocouple at half radial distance gives about twenty
r~'
r,
r\,
(~ percent error in thermal diffusivity values due to sharper
t\ a, CD o co
CD t\
temperature profiles. Onthe other hand, the temperature
,~~, c (Y)

ao 'Y)
~ a), th,
ul) (~
n
(o oo

i
'~ (~
o profile near centre is almost flatter minimising such
'J:) ch'
ht
O
h~P n
'
A4mjjl~ '
'~'
h c~
'

Moreover, it is not desirable to insert two

=D.
"
D"
'~ ' l
s*_ **
h'

~ ~1
~
~Jj~4M'4!~f~~~

40. u"
~h/}~/rlY

__
errors.
thermocouples in pellet body specially whenthe effect of
29' bulk density and porosity needs to be investigated. In
Fig. l. x-ray diffraction pattern ofiron ore, raw and fired. view of the above facts, the surface and centre tem-
(a) raw ore perature were measured and results analysed by ap-
(b) fired at 1223 K plying principles of heat balance.
(c) fired at 1473 K
In the present setup the pellet is fixed between three
indicates the chemical analysis of the ore. The colour of chromel-aiumel thermocouples so that one thermocouple
the ore was orange brown. The X-ray analysis of this penetrates the pellet through drilled hole with its bead
at the centre. The other two thermocouples are diagonally
ore powderwascarried out in Rigaku-D. Max111 X-ray
opposite to the pellet diameter with their beads touching
diffractometer. The analysis (Fig. l) indicates the
the surface. The insulators are inserted to the wires
presence of goethite (Fe203 H20) along with hematite •

in the ore. The peaks at 'd' values of 4.189, 2.449 and keeping certain distance from pellet surface. The short
circuit between wires is avoided by coating the wire with
at I
.717 represent goethite corresponding to crystal
thin layer of iron oxide (slurry/dried) and maintaining
planes of (1 lO), (1 1) and (221) respectively. The peaks 1 safe distance between two wires. Loose iron ore powder
at 'd' values of 3.689, 2.701, 2.52, 2.207, 1.84 are of
hematite. 12) Thedensity of the raw powderand its density was filled to avoid any air gap between thermocouple
and pellet hole. Both the outer thermocouples were tled
after firing at higher temperatures, were measured by
specificgravity bottle using kerosene as medium. to a thin walled hollow stainless steel tube containing
Pure iron oxide wasprepared by precipitation of ferric centre thermocouple and clamped. (Fig. 2)
hydroxide from analytical grade ferric chloride using An independently controlled, preheated tube furnace
is lowered on the pellet and the rise in surface and
ammonium hydroxide. The precipitated hydroxide was
centre temperature of the pellet were recorded by two
washed, dried and then fired at 273 for 90min to 1 K
decompose the hydroxide. The completeness of de- pen strip chart recorder at a chart speed of lOcm/min.
confirmed by X-ray analysis. Relationship between surface temperature and centre
composition is
temperature was also recorded in an omniscribe x-y
2.2. Preparation of Pellets recorder with accurate scale (lcm I mv). As tube
=
The ore powder was pelletised in a disc pelletizer of furnace does not produce an isothermal reservoir but
O.6 diameter and O. 15 lip height. Onepercent dextrin establishes a temperature plateau some where at the
m m
was used as binder. The overall rolling time was about centre of the tube, its design is modified, by using
90 min. The green pellets of size lOl 6 and spherical reflectors above and below pellet location. These
shape were picked, dried at 423473K for 2hr and then
mm reflectors made of stainless steel are almost of same
fired at 1223 K for I hr to attain workable strength of diameter as the furnace tube. Their use decreases the
pellet. These pellets
were furnace cooled. The weight and convective air inspiration, if any, from the bottom of the
volume were measuredto calculate their bulk densities furnace and also provide configuration similar to an
and porosities. The true density of fine powder was enclosed chamber. It also increases the length of hot zone
measured by density bottle using kerosene as mediurn. and variation of temperature within it is decreased. This
The pellets of pure iron oxide were madeby prolonged is also helpful in maintaining the pellet position at the
hand rolling and then on glass plate. centre of furnace tube. The direct radiation from furnace
A radial hole of O.1 cmdiameter was drilled upto the wall to thermocouple bead was avoided by placing a
centre of pellets and classified according to their po- dome shaped ceramic/metallic shield on the lower
rosities. The pellets were subsequently fired at temper- reflector to envelop the pellet. This decreases the lateral

C 1996 ISIJ 992

 IS[J International, Vol. 36 (1 996), No, 8
temperature gradient and hence a]lows to measure indicated in Fig. 4.
surface temperature more accurately. The heat flux Attempts were madeto correlate 'T~' and 'T.' with
passing through the thermocouple wire is likely to cause time 't'. Out of the expressions attempted, the expres-
someerror in evaluation of thermal diffusivity. Assuming sion of the type
the thermal contact between the pellet and wire to be
ekt) + T~e~kt
sufficiently good to produce similar temperature gradi-
T.
= T~(1 - ..........(1)

ent, it was estimated that error of about 2.50/0 in ther- was found to suit best, 'T~' is the initial
temperature of
mal diffusivity value.11) Sameorder of error
may be pellet and 'k' a parameter. Ray
is et al.13) derived such
anticipated in the present case. an equation assuming exponential rate of heating of
volume element of sample.
3. Results and Discussion Validity of the Eq. (1) in the present work was tested
by plotting a curve between Iog.((T~- T.)1(T~ T.))
Figure 3 indicates the rise in surface (average of two -
with time and represented in Fig. 5 for two sets of
-
thermocouples) and centre temperature of a pellet with
readings. The linear plot confirms the suitabiiity of the
time for one of the sets. The relationship between sur-
expresslon for most of the region of heating, except at
face temperature (T~) and centre temperature (T*) is
final stages when' T.' approaches ' T~' and the curve tends
to be asymptotic. Asimilar type of expression correlating
L
I
temperature of a volume element at a radial distance 'r'
!1 can be envisaged as
lI
T~(1 -e~k't)
1Graphite 810ck T.= + T.e~k't ..........(2)

2 2Furnace Tube where k' another parameter.

is

3 3Sta inless St ee
Rod.
[ Relation between and k' is necessary for further k
analysis of the results and needs to be evaluated.
4Refiecio r Taking
4 5Pe etil

6Thermocoup[e
k' =k(R/(R-r))" .....
..........(3)
11

7 ,,
where 'R' is the radius of the pellet, Eq. (2) modifies to
5 8Insulator
7 6 7 T.
= T~( I - e~
kt(R/(R - '))")
+ t.e ~ kt(R/(R - '))" .(4)
9Ref[ecto r
. . . .

10StainlessStee[ The above equation our boundary conditions

satisfies
9 Tube for any value of The value of 'n' in the present case
'n'.
8 8 8 determined experimentally. The temperature at
lo 11Shield was
surface and the half radial distance r R/2 were measured
=
during heating of a pure iron oxide pellet whenexposed
to furnace. Subsequently for the samepellet, a hole of
half radius was extended to centre and 'T~' and 'T.'
were recorded. Three runs were taken to check repro-
ducibility in each case. Figure represents the variation 6
of temperature at surface, half radius and at centre of the
iron oxide pellet with time. The parameter k' and were k
Fig. 2. Location of pellet, thermo couples, reflectors and evaluated by plotting -10g.((Ts T./2)1(Ts T~)) and
shield. ~for ~
-loge((Ts~T.)/(T.-T~)) vs. tirne this pellet. The

l073

973

873

773
Q; Ts
*
::,

15 673
*g,
E 573
1,

Tc
Pe[tet Fired ut 1303 K
Size = 0•6921 cm.
473

373
Fig. 3.
Variation of surface and centre temperature of pellet
15 30 75 150 with time.
45 60 90 105 120 135 165 180
Time (s)

993 C 1996 ISIJ

 ISIJ International, Vol. 36 (1996). No. 8
results are indicated in Fig. 7. The ratio between the The average temperature of pellet at an instant was
slopes of these two linear plot k'/k was found to be .238 I further calculated by using an expressionl4)
and the value of 'n' in Eq. (3) as 0.3.
Thus the expression modifies to 4'rcr2Trdr
kt(R/(R-'))0.3) T.e~kt(R/(R '))o '
T.= Ts(1 e + .,..(5) Tav~ .
(7)
4lcr2dr
on simplification
To)e~kt(R/(R -'))0.3 (6)
(T=
T.
= T~- - .. . ... . ... and substituting the value of ' T.' from Eq. (6) and further
Rayl3) reported that the parameter 'k' is not a constant integration.
but may vary with process conditions. In view of the The calculated value of ' T*+' with time is plotted in
above the temperature ' T.' is calculated with instantan- Fig. 9 for two sets of experiment and other values are
indicated in Table 3.
eous 'k' value at any time. The temperature of elements
of I thickness and with radial distance 'r' equal to
mm
(R-0.05), (R-0.15), (R-0.25)• is calculated using • •
3.1. Calculation of Thermal Diffusivity by Heat Balance
The heat balance at the surface of pellet dictages that
expression (6) for various time intervals. The results for
amountof heat passing through the pellet surface inside
one of the set is indicated in Fig. 8, which in fact, indicate is utilised in heating the pellet.
the temperature profile within pellet at various intervals.
Thus,

dT dT..
973 K' A -

dr = V'B• C•
dt "'
..........(8)

where, K: thermal conductivity of pellet,

S73
V: volume,
B: density,
773 C: heat capacity, and
X
Q,
A: surface area of pellet.
:~ Rearranging Eq. (8)
rg
6
Q,
73
dT.~
a;
E =(K•A/(V•B• C))•(dTldr) ............,..(9)
OJ
dt
H
aJ
U 573
,~ substituting for 'oe'
1:
:,
U1
dTav
=(30c/R)•(dT/dr) ..........(10)
4 73
dt ......,,

x Pe[[et Fired at 1303K Size 07863em. The thermal diffusivity

can be evaluated for known
(oc)
373 values of (dT*./dt)
1523K 0•5854cm.
and (dT/dr). The value of (dT*+/dt)
. ," ,,

for various time interval, can be evaluated from 'Tav'

values indicated in Table 3 and Fig, 9. Evaluation of
373 473 573 873 773 873
273 dT/dr is difficult and require certain assumptions. Many
Centre Te~perature (K)
expressions relating dT/dr with heat transfer parameter
Fig. 4. Relationship between surface and centre temperatures.

o -
PetletsFire(1 at 1473K Slze 0636cm
o
Ff
F;u' x - " 1303K ," 0~92 ,,
~2
\ ~
u'

~
1

l Fig. 5.
oD 15 30 45 eo 75 90 105 120 135 Iso 165 Validity of Eq, (1).

l~&~:)

C 1996 ISIJ 994

 ISIJ Intcrnational, Vol. 36 (1 996), No. 8
Ts ~:
1173 973 u'_
F~~ ~
E
F: l
1073
873
q
150
t
973
~;' 13S
,
873 o 773 ,
o
~
~( oTrl2 120
a'
*:) 773 Tc , 1a5
~;
*Q' o
o
~!(
673
673 ,
o 90

o Pure F~203 ru
573
573 Q'o~

Temperature= I 008 K 7S

473
Pe[tet Slze =1•310cm,
E
~:l
I473 60

373
45

273 373
30
O 30 60 90 120 150 180
liEDgi~) 15

273.6:15
Fig. 6. Variation of surface, half radial and of
at centre ' 515 •415 '31; '215 •115
o
temperature of pellet with time.

Fig. 8. Temperature profile and movementof heat front

0,9(T*- T.) point.
3.5
Temperature = I 008 K
Pel[etSize =1,310cm. o
3.0 T=Tr/2 973
Fo
I
2.5
873
~u'
T=Tc , I
Pel et Fired at 1473
K Size 06154cm
\
1: 20 X " 1303K ,, 7863 ,,

,L 773
~, 1.5
o' ~(
o
,~}~!'>
673
1.O
'l'

o :~

D•5 fl'

L
Q' 573
HE~

o 30 60 90 120 150 IS
O
Time(s) 473
Fig. 7. Evaluation of n by ratio of slopes kk
373
'o(' for different boundary conditions are available in
literature.15) These expressions are not useful in present
case as the basic aim is to evaluate 'oc' through dT/dr. 273 15 30 4 SO 7S 90 10S 120 135 150 165 Ieo

Figure 8
indicates that the thermal profile within the Ti
me(s )
pellet could be approximated to linear except at surface Fig. 9. Average temperature rise of pellet.

and initial stages of heating. It is also evident that the

depth of major temperature fall from T~ to T*, change in Fig. I along with X-ray peaks of raw ore powder. The
with time of exposure. Initially the heat front is con- phase analysis indicates that goethite peak (d=4.189)
centrated at narrow depth near surface and with time it disappears due to its decomposition which is expected
penetrates inside. In present study 90 o/o of temperature to occur8) between (300-500'C) (573-773 K). Newpeaks
fall from T~ to T* is assumedat major change and the at 'd' values of 407 298 278 244 and 2.03 matches
depth at which occurs is calculated by Eq. (5) to
it with 20, 100, 40, and 50 "/o intensity (1/I~) peaks of
calculate dTldr. Onesuch model calculation is shownin p-CaSi03 (Triclinic).12) The equilibrium diagram16) of
Table 4. Theaverage of best (7-8) readings are considered CaO-Si02 system also indicate the possibility of this
to calculate average thermal diffusivity. phase at the temperature and compositions under
The X-ray diffaction analysis of the powder samples considerations. Laybidil7) and co-workers working on
K
heated at 1223 (950'C) and 473 (1 200'C) are given 1 K Kudremukhiron ore observed the mineralogical changes

995 C 1996 ISIJ

 ISIJ lnternational, Vol. 36 (1996), No. 8
Table 3. Average temperature of pellet (Data rounded off to nearest integer).

Time I .6921 .6839 .6249 ,5854 .6146 .6154 .6921 .636 .6294 .6154 .7863

(s) 2 28 12 2914 31 54 3521 3650 3031 2812 3057 3154 303 1 2654
3 1223 l 393 1488 l 523 l 548 1473 l 303 1473 1488 1473 l 303
15 309 311 323 330 320 313 316 323 321 307
30 355 362 377 372 400 380 358 367 369 382 356
45 418 43 l 449 448 474 457 445 45 l 436 46 l 418
60 489 578 537 534 553 54 l 530 518 513 544 492
75 563 610 622 620 633 626 620 603 593 625 569
90 634 692 698 702 703 702 696 680 664 702 639
l05 698 760 760 768 757 770 763 746 737 766 702
l 20 749 817 821 826 819 823 808 800 793 813 756
135 793 865 863 868 855 868 857 845 835 854 801
l 50 829 897 904 908 89 l 903 892 885 877 89 1 837
165 862 919 926 939 920 919 915 913 92 l 872
l 80 886 935 958 938 937 939 942 895

l: Size in cm, 2: Density kg/m3, 3: Temperature K,

Table 4. Calculation of thermal diffusivity. Pellet size 0.6291 cm, fired at 1303 K, T~=293 K,
T* Tc Radial c(
Time
(K) (K)
Ts
~ Te Ts
~ To 0.9(T.
-
T.)
distance dTld r dT**/d t (m2/s)

15 358,5 306 52.5 65.5 47.25 0.1889362 250.08442 2.

1667 l.998 x lO~ 7
30 442 366,5 105.5 149 94.95 O.2848327 333.35357 4.40 15 3.046 x lO~ 7
45 573 399 174 280 l56.60 0.3552634 440.79969 5 .7406 3.004 x lO~?
60 665 465 200 372 l80.00 0,4103466 438.65356 5.8242 3.063 x l0~7
75 745 540.5 204.5 452 184.05 0.4506661 403.03410 5.5318 3.166 x l0~7
90 805 610 195 512 175.50 O,4900077 358.15763 4.7697 3.072 x l0~7
l05 851 677 174 558 156.60 0.5786412 30 1.94284 3,7425 2.859 x l0~7
l 20 881 726 155 588 139.50 0.5371069 259.72483 I
3, 176 2.769 x l0~7
135 910.5 783 127.5 617.5 l 19.75 0.5579178 205.67545 2,7725 3.1 10 x l0~7
l50 930 830 100 637 90.00 0.5754939 156.38740 2.0873 3.079 > l0~7
165 946 870 76 653 68.40 0.5901223 ll5,90817 l .5467 3.079 x l0~7

(Average: 3.025 x 10 7).

during firing and reduction. They observed the formation
of fayalite bridge between iron ore particles which melts
and covers the iron oxide. In the present case no fayalite s,o

was observed in X-ray analysis of fired ore. Wynnyckyj

and Fahidyl 8) reported that trivalent iron oxide does not
4.0
form iron silicate under oxidising conditions, which is in
confirmation with our observation. Yankand Standish i 9)
opined that iron ore sintering is basically a liquid phase ::~

sintering process due to the formation of low melting u' 3.0

ferrites in presence of basic oxides as CaO. Nasr20) and
8
J
co-workers also reports the formation of calcium ferrites J:
~1
"I' 2.0
with their studies on Beharia iron ore pellets. X-ray ~
analysis (Figs. 1(b) and l(c)) does not indicate the
presence of ferrites either, in the present case, probably 1.0
due to low CaO/Si02 ratio. The p-CaO Si02 formed in •

induration of pellets is fairly high melting silicate and

formation of liquid phase is not possible under prevailing O
conditions. In view of above, the present case is a solid 573 673 773 873 973 1073 1173 1273 1373 1473
state sintering process accompanying mineralogical Temperature of Aring K) (
changes.
Fig. 10. Weight loss of iron oxide on firing.
The weight of the iron ore powder on heating
loss
is indicated in Fig. lO. The total weight loss is around
4.20/0 and mainly occurs below I 173K (900'C). The the ore.
decomposition of goethite is expected to be completed The decomposition of goethite and other gangue
K
around 773 (500'C) or s08) and hence the continued minerals, formation of p-CaSi03 and probably the
weight loss upto I 173 K (900'C) maybe attributed to sintering of intraparticle pores, especially at higher
decomposition of other gangueminerals associated with temperature change the density of the powder. These

C 1996 ISIJ 996

 ISIJ International, Vol. 36 (1996). No. 8
Table 5. Variation ofthermal diffusivity/conductivity with density/porosity.

Weight of Theoritical Ratio of

pellet
Size Volume Density Porosity e(
K diameter speciiic Treatment
(cm) (cm3) (m3lkg) (~/*) (m2/s) (W/m ' s)
(g) 7.ero poro. volume

3.9043 0.6921 l .3886572 2812 35,36 3.062 x lO7 0.9015 0.5983956 l.5471820 Fired at 1233K( 950*C)
3.9043 0.6921 l .3886572 2812 35,36 3.025 x l0~7 0.8903 0.5983956 l.5471820 Fired at 1303 K (1 030'C)
5.4500 0.7863 2.0363543 2654 38,48 2.9875 x lO~ 7 0.8301 0.6687640 l.6253472 Fired at 1303 K (1 030"C)
3,9043 0.6839 l .33988 12 29 1
4 35,24 3.016 x 10~7 O.920,_ 0.5916715 l .5443 144 Fired at 1393 K (1 120"C)
2.9585 0.61535 0.9760 119 303 l 32 64 3.25 x l0~7 I.0337 0.5394 154 l .4845563 Fired at 1473 K (1 200'C)
3.2939 O 6360 l .0776056 3057 32.07 2.94 x l0~7 O9409 0.5590745 l .472 1846 Fired at 1473 K (1 200"C)
32939 0.6294 l .0444036 3154 33.46 2.809 x lO~ ? 0.9279 0.5494747 l 5029232
.
Fired at 1488 K (1 215'C)
2.9585 0.5854 0.8403,_36 3521 25.72 2.876 x lO~ 7 l .0602 0.5301532 1.3963375 Fired at 1523 K (1 250'C)
3.5496 0.6283 l .0389380 34 1 6 27.93 2.9517 x l0~7 l .0557 0.5633402 1 .3873592 Fired at 1523 K (1 250'C)
3.5496 0.6146 0.9724476 3650 24.27 3.025 x l0~7 l 1560
.
0.560206 1 l .3204875 Fired at 1548 K (1 275"C)
3.2939 0.6294 1.0444036 3154 33,46 2.7425 x l0~7 0.9056 0.5494747 l 5029232
.
Fired at 1488 K (1 215"C)
2.9505 1
0.6 535 O.9760119 3031 32,64 3.045 x l0~7 0.9668 0.5394 154 l 4045563 Fired at 1473 K (1 200"C)

Table 6. Variation of thenTlal diffusivity/conductivity with density/porosity (Pure iron oxide)

Weight of Theoritical Ratio of

pellet
Size Volume Density Porosity C(
K diameter specific Treatment
(g) (cm) (cm3 ) (m3/kg) ("/~) (lll2/s)
(W/m ' s)
zero poro, volume

,_.6393 0.5756 0.7988234 3304 30.44 2.756 > lO 7 0.9534 0.5099983 143766 Fired at 1523 K (1 250'C)
2.6393 0.5756 O7988234 3304 30.44 2.782 x lO~ 7 0.9624 0.5099983 l .43766 Fired at 1523 K (1 250'C)
3.7530 O6690 1.2542004 2923 38.46 2.50 x lO~ 7 0.765 l 0.5734982 l 58739
.
Fired at 1473 K (1 200'C)
3.7530 0.6490 11450451
.
3278 30.99 2.649 x IO~7 0.9074 0.573498•_ 1.44923 Fired at 1498 K (1 225'C)
2.6393 0.5672 0.7643585 3453 27.3 1 2.704 x lO~ 7 0.9776 0.5099983 l .37563 Fired at 1598 K (1 275'C)
True density of lron Oxide Powder4750 kg/m3

density values were taken as true densities to calculate

porosity of pellet after firing at appropriate temperatures.
(d) K./K=
= (1 - 2e( P- l)/(2P + I))/
(1 +e(P- 1)/(2P+ 1)) ..........(14)
The measureddensity values are .......

K./K, (c'2/3 e2/3))/

Temperature Density (kg/m3)
(e)
= + p(1 -
(e2/3_e+p(1 -e2/3 +e)) (15)
l 223l 303 K 4350 .................

l 393l 473 K 4500 where, K. : thermal conductivity of pellet with zero

l 3881 523 K 4740 porosity,
l 448 K 4820 Kg : thermal conductivity
of gas in pore,
Table 6 indicates the
values of thermal diffusivities : P
K~1Kg.
Figure ll indicates the experimental points of thermal
with pellet density, porosity, and corresponding thermal
conductivities with porosity of the pellets along with
conductivities alongwith the thermal treatment given.
calculated values of effective thermal conductivity by Eqs.
The thermal conductivlties are calculated by taking21)
(14) and (15). The variation of 'K*' with porosity
the true heat capacity value of iron oxlde at 973 K(700'C)
computedby Eq. (1 l) is too low and that by Eq. (13) is
as I .047 kJ/kg K.21)
too sharp to explain present experimental values.
3.2. Model Fitting However, it could be noticed from the figure that values
A simple expression for effective thermal conductivity computedby Eqs. (14) and (15) fits in satisfactorily for
of iron ore pellet to its porosity is useful in solving the 'K*' values of .55 to I I
.60W'm~ I s~ I for Eqs. (14) and '

heat transfer problems. Manyexpressions used for iron (1 5), repectively. Thethermal conductivity of air for these
calculations were taken22) as O.06W'm~1
ore agglomerates are listed by Akiyama8) and co-workers. 's~1. Equa-
They explained thelr resLrlts by applying unit cell model tion (14) is a better fit out of the two equations. It could
originally proposed by Luikov. In the present invest- also be observed that the present experimental values of
igation attempts were madeto correlate the effective effective thermal conductivities increases more sharply
thermal conducti\'ity 'K.' and porosity 'e' by different with the decrease in porosity than those predicted by
model equations availabie. The radiative heat transfer these mode]s. Thls aspect motivated us to analyse the
within pore is neglected. The equations tried are results further by using basic sinterlng equation available
in literature.
(a) K.=e'Kg+(1 -e)K~ ....................................(1 1)
The process of slntering of fine powders were widely
(b) K.=[e/Kg+(1 -e)/K~]~1 ............................(12) investigated in powder metallurgy. Various mechanisms

were proposed and activation energies calculated.23,24)

(c) K*=K~~••K~ •••
••••••••••(13)
The sintering behaviour of iron oxide in specific were

997 C 1996 ISIJ

 ISIJ lnternational, Vol. 36 (1 996). No. 8
reported by Misra,25) Seshadri26) and Wynnyckyj. 18) The iron ore pellets and pure iron oxide pellets are indicated
rep.resentation of densification of pellet with parameters in Fig. The extrapolation of the plot to ratio of '1'
12.
which approach '1' on sintering to theoretical density will lead to the value of 'K~'. The logllog plot for same
results is indicated in Fig. 13 and the intercept values are
were used to interpret sintering kinetics. Use of volume
ratios by Tikkanen27) to explain sintering of metal calculated by least square method for better accuracy.
powders and by Wynnyckyjl8) for iron ore pellets are The values of 'K~' thus found for iron ore is
reported in literature. Wynnyckyjl8) even co-related the
crushing strength of pellet to volume shrinkage ratio and
l.81
W m~I s~ I at 973 Kand 1.90 •m~I s~ I for pure
• '

iron oxide pellets. Limited numberof points, in case of

W '

to percent theoretical density of pellet. With the same pure oxide is due to less quantity of synthesised material
10gic, an attempt has been madein present investigation available. The pellets were hand rolled in this case.
to correlate the thermal conductivity to ratio of volume
of pellet to its volume at zero porosity. The results for 4. Conclusion
(1) Theiron ore from Rajhara mine oflndia contains
goethite alongwith hematite. Onfiring goethite decom-
1•8
pose and on reaction p-CaOSi02is formed.
(2) The decomposition of goethite and other gangue
1•7 minerals along with formation of CaOSi02changes the
equation (1
5)
1•6 eqvation (i4 )
for Ks=1•6 w.m~1~1
1•5
for Ks=1•55 ,, 1•5
\
'i for Ks=1•5 ,,

~ \
1•4 1•4 lron ore
X
o Pure lron Oxide

~
~
1.3
1.2
1•3

1.2
~
~1 .1
\
"~~'\ 1•l
~
* 1•O 1•O
o
~)(~\
\ c X
0•9 X\ \ +\ 0•9 X
~\~~

oe O'
s
0•7
30 35 40
o 5 10 IS 20 25
EgrgSi!tyl~~)
1.2
1.3 14
•
1•5 1•6
(V.actual )/( V, theorit ica[
)
Fig. 11. Variation of thermal conductivity with porosity and
Fig. 12. Variation ofthermal conductivity with volume ratio.
model fitting.

x:
8 *
~
\
75 2
_~ X
X lron ore 18
X
X
o Pure lron Oxide 16
eD

1~
o
12

10
ll
ll 8
6
4
2 Fig. 13.
Log-log plot between thermal conductivity and
52 volume ratio.
28 26 24 22 20 18 16 14 12 10 2 O-2-4
8 6 4 wrnlS1) -6 -S -10-12 -14xl

[og (K

C 1996 ISIJ 998

 ISIJ International. Vol. 36 (1 996). No, 8
density of powder. NaukUSSR,Metally (1976), No. 6, lO7; Russian Mela!!., (1976),
(3) The thermal diffusivity and
thermal effective No. 6, 9-15.
conductivity of iron oxide pellets for radial heat flow can 1O) A. Watts, J. K. Wright and J. Vale: Trans. Inst. Min. Metall., 89
(1980), cl92.
be determined by calculating average heating rate and
applying heat balance. Theresult are comparable to those l l) A. Watts and J. K. Wright: T,'ans. Inst. Min. Metal!.. 91 (1982),
C18.
given in literature.
12) Selected Powder Diffraction Data. JCPDS,Card No. 13534,
(4) The decrease in effective thermal conductivity 19-249, 17-536.
with increase in porosity could be explained satisfactorily 13) H. S. Ray: Proc. Symp.Advancesin Chemical Metallurgy DAE,
by model Eqs. (14) and (15) in the text. Bombay,(1979), 43/1 1/18.
(5) The effective thermal conductivity at zero
14) J. Szekely and N. J. Themelis: Rate Phenomena in Process
Metallurgy. Wiley Interscience. NewYork, (1971), 203.
porosity of the pellet could be determined more 15) H. S. Carslow and J. C. Jaeger: Conduction of Heat in Solids,
accurately by using basic shrinkage parameters and Oxford University Press, London, (1971), 203.
densification. l 6) The Making Shaping and Treating of Steel, ed. by H. E.
(6) Thevalue thermal conductivity at zero
ofeffective McGannon,USS, Pittsburg, (1964), 304.
l 7) J. I. A. K. Bhattacharya and Prakash: Proc. Advances
Laybidi,
porosity iron ore pellet is

iron oxide pellet is around I .9 W•m~I s~ 1.

W
I .81 m~I s~ I and for pure
• '

'
in Chemical Metallurgy, ed, by C. K. Gupta and M, K. Totlani,
I.1.M. Material Div.. Bombay,(1991), 325.
l 8) J R, Wynnyckyj and T. Z. Fahidy: Metal/. Trans., 5 (1974),
99 1
REFERENCES .

l 9) Y. H. Yangand N. Standish: ISIJ Int., 31 (1991), 468.

1) J. R. Wynnyckyj and W. A. McCurdy: Meta!1. Trans,, (1974), 20) M. I, Nasr. A. A. Omarand M. E. H. Shalabi: J. Inst, Eng,
2207, (India), 75 (1994), 79.
2) E. C. Seaton, J. S. Foster and J. Velasco: Trans. Iron Steel Inst. 21) E. I. Kazantsev: Industrial Furnaces, Mir Pub., Moscow,(1977),
Jpn., 23 (1983), 490.
3) S. K. Misra and V. K. Jain: J. Inst. Eng., (India),
i 24 .

Metall. Mater. 22) R. T. L. Guthrie: Engineering in Process Metallurgy. Clarendon

Sci.,67 (1986), 17. Press, Oxford, (1989).
4) I. Campbell and E. M. Sherwood: High Temperature Materials
E. 23) F. Thummler and W. Thomma:Metals and Materials and
and Technology, John Wiley, NewYork, (1967), 960. Metallurgical Review, (1967), 69.
5) H. Ohta, T. Akiyama, I. Suh, R. Takahashi, J. Yagi and Y. 24) A. Mohan. N. C. Soni and V. K. Moorthy: T,'a,Is. Inclian Inst.
Waseda: Tetsu-to-Hagan~, 75 (1989), 1877. Me!., 33 (1980), 423.
6) T. Akiyama. G. Ogura, H. Ohta. R. Takahashi and J. Yagi: 25) U. N. Mishra and P. K. Sen: T,'ans Indian Insl. Met., 32 (1979),
Tetsu-to-Hagan~, 77 (1991), 23 1 .
41 l.
7) T. Akiyama, R. Takahashi and J. Yagi: ISIJ Int., 31 (1991), 24. 26) G, L. Misra. A. R. Das and V. Seshadri: Trans. Iron Stee! Inst.
8) T. Akiyama. H. Ohta. R. Takahashi, Y. Wasedaand J. Yagi: Jpn., Il (1971), 323.
ISIJ Inl., 32 (1992), 829. 27) H. F. Exner and G. Petzow: Sintering Processes. Material Sc.
9) V. M. Abzalov. B. P, Yuryev and S. G, Bratchikov: IZV Akad Research VoL 13, ed, by G. C. Kuczynski, I 13.

999 @1996 ISIJ