water in the microwaveoven to boiling. 2.3.2. Reduction ... free water and then bound water. Since all samples ... by a further heating at increased temperatures. It is sufficient to report here that the weight per ... moisture removed before COreduction commenced ... 400'C and the attack becameuniformly extended throut) ...
ISI
Reduction of Microwave lrradiated N.
J International,
Vol.
ore Particles
lron
31 (1991), No
in
I
pp,
Il-16
CO
and PRAMUSANTO1) STANDISH 2500, NSW
Department of Materials Engineering, The University of Wollongong, Wollongong, Development Centre, Jl. Jenderal Sudirman 623, Bandung, Indonesia. (Received on February
19,
1990, accepted in the
form on July 20,
final
Australia.
Mineral Technology
1)
1990)
The effect of microwave (2.45 GHz)pre-treatment on the reduction of magneto-hematite ore particles in COwasinvestigated. Itwas found thatthe presence of magnetite in the ore has considerable effect on the microwavetreatment and that the oxidation of the goethite, present in the ore, is obtained after the magneto-hematite grains are almost entirely oxidized to hematite. The results of the reduction tests showthat microwave treatment enhances reduction but benefit optima exist. Inthe present ore optimumreduction wasobtained for an irradiation of min at a powerlevel of 1300 W. It is also shownthat in gain post-reduction strength and decrepitation were likewise improved by miorowavetreatment. addition to a reducibility
6
microwave treatment; KEYWORDS:
1
magneto-hematite;
COreduction.
Introduction
.
As is well known, the reduction ofiron
ore
is
great-
ly enhancedwhen the ore is in the form of sinter or pellets which provide a porous structure needed for
gas-solid contact. However, sinterintg and pelletizing operations involve high capital cost installations and are not generally economic in small voleffective
umeironmaking. Onepossible methodof enhancing reducibility that mayprovide an answer in small scale ironmaking is
the microwave precursor treatment method. It has been recently shownl) that microwave irradiation of minerals, including iron ores, reduced energy consumption in subsequent grinding because the microwaveprecursor treatment caused thermal stress cracking of the ore particles. However, the potential effect of such microwaving of ore lumps on their reduction behaviour has not been investigated. The purpose of this paper therefore is to present the results of an investigation of the effect of microon the reduction of a magnetowave irradiation hematite ore as an alternative to sintering or pelletizingr, for a small scale charcoal blast particularly furnace ironmaking in Lampung,Indonesia. It is considered that such an investigation would also yield someuseful information in the general field of microwave applications, quite apart from providing
immediate data or
pelletizint)cr
Experimental
2. 2.
for a possible alternative to sintering needs referred to above.
for the
Reduction tests were carried out in the standard thermogravimetric apparatus provided with COand N2 gas control, flowmeters and data logger interfaced with a computer. Materials
2.2.
Magneto-hematite ore particles (13-15 mm)prepared from Lampung lump ore were used in the investigation. The chemical composition of the ore is given in Table
I
.
The measuredaverage
density and porosity of the 3.94 g/cm3 lumps and 16.2 o/o' respectively. ore were All samples were dried in a conventional oven for 24 h before testing.
Microwave drying was not used because of the
JV:.B.
possible effects
on subsequent results.2)
2 3. Procedure 23 I Microwaving A weighed test sample was put .
.
.
.
in a ceramic cru-
which was then placed onto an insulating brick located in the centre of the oven. Full power was turned on and the sample irradiated for the required time. After irradiation, the sample was cooled in the air and weighed. Several samples were also used fbr temperature measurementduring microwaving. For this purpose a stainless steel sheathed Chromel-Alumel thermocouple was inserted through the roof'of the cavity into cible
a hole drilled in the sample. The stainless steel sheath was grounded to the oven or the chart recorder due to microwave field. The to prevent fluctuation
I Appara tus .
Microwave heatint:g was carried out at a frequency of' 2.45 C.Hz in two power levels, viz 700 and 1300 using SHARP Models R-9360 and R-2370 microwave
W
ovens, respectively.
C 1991 ISI J
Table FetoL:ll
66
.
8
l.
Fe +
5 80
Chemical composition of Lampungore. Si02 Al203 O.3
0.6
CaO Ol
Ti02 .
02
(o/o)
P205
Mn
o 06
o 66
,
,
ll
ISIJ International,
Vol.
31 (1991), No, l
thermocouple was calibrated by bringring a beaker of water in the microwave oven to boiling.
1ooo
Reduction
2.3.2.
The tube
furnace of the thermogravimetric apparatus was switched on and heated in the presence of N2 flow to a constant temperature. single particle sample was then suspended into the tube of the furnace from a balance in the usua] way and the logger switched on. Whenthe weights of the sample becameconstant N2 was replaced by and the reduction commenced. After a prescribed reduction degree was reached, was again replaced by N2 and the furnace switched off. After cooling in the furnace to the N2 temperature, room was turned off and the sample was taken out and used for microscopic and other tests. The reduction degree was calculated as the loss of oxygen from the initial total iron oxide obtained from the chemical analysis (Table ).
A
CO
CO
800
('
600
~;
*
~} ~$
N0.5 A;
1,300 w
,
N0.5 B;
1,300 w
l
N0.5 C;
700
w
o
N0.5 D;
700
w
o
N0.51~;
700w
E:
~~
~
400
200
o o
I
1
2
3
4
5
6
7
8
g
10
11
Time, min
Results
3.
Fig,
Microwave heating
l.
Microz~)ave Heating Rates
3. 1.
The results
of microwave heating of the
ore samples are Fig.
shownin
I shows that
remarkably stabilised
therefore
Fig.
the initial
1ooo
Lampung
c
1300 Wwas
than at 700 before temperatures at approximately 800'C. This result is in line with the previous finding of McGill
et al.3)
800
U
A comparison hematite
ore from Whyalla, South Australia, was also tested and the heating rates, shown in Fig. 2, were unexpectedly high. The reason in this case oxide-a was the presence of
manganese
microwave hyperactive material-as reported previously by Ostwald4) and also confirmed here by
soo
g' :$
c~
o 0
~ H e)
400
XRDanalysis.
200
Weight Loss on Micro~)aving In the absence of chemical reactions weight loss on microwave heatinbcr is associated with the removal of water-at first free water and then bound water. Since all samples were oven dried before being used for tests, only the bound water would be removable by a further heating at increased temperatures. It is sufficient to report here that the weight per cent removed under all conditions of microwave power and treatment times used was 0,4-1.1 O/o' and that the value of I I o/o was also the amountremoved on heating in a nitrogen atmosphere without prior microwave treatment. In other words, the total moisture removed before reduction commenced was the samein all cases. 3.2.
.
o o
Microscopic Observation
Recent data reported by Walkiewicz et al.1) demonstrated that the stresses generated by microwave heating' depended on the heating rate as well as on the Consequently, rapid and setemperature attained. lective tial
12
heatintg of ores maygive the required differenstressing at relatively low temperatures also.
1
2
3
700W
E]
Whyallaore;
e
Whyalla ore : 1300
4
5
7
6
W
8
9
Io
11
Time, min Fig.
2.
Microwave heating
at
700 and
1300 W.
Whyalla
ore.
Fl,g
ples l)
2)
CO
3. 3.
1 300 W. Lampung
1.
heating at
W
faster
at 700 and
ore.
3)
3 shows the
results
of heating the present sam-
In the muffle furnace in air for
I h at:
400'C (A-1); 800'C (A-2); I OOO'C(A-3). In the microwave oven at 700 : I min (B-1); 2min (B-2); 4 min (B-3). In the microwave oven at 1300 W: 3min (C-1); 6min (C-2); 10 min (C-3).
W
The structure of the Lampungore5) consists of hematite and magnetite phases as well as cavities and intergranular gaps fllled with finely layered goethite. It is often found that the hematite envelops goethite in nodular-like
structure.
Onconventional goethite
heating (Fig, 3(A)) it is seen that into hematite-like transformed pattern was
ISIJ International,
Fig.
3.
Microstructure
of'
I.ampung ore
Vol.
after
and magnetite grrains were oxidized to hematite. The were apparently attacked from the edge at 400'C and the attack became uniformly extended
31 (1991), No, l
various heating regimes (see text for details).
Table
throut)crhout
the grain at
I OOOoC. Cracks
appeared
800 and I OOO'C. With microwave heating it appears that the magnetite grains were oxidized before goethite. Cracks were observed after 2and 4min of microwave heatingr and more obviously after 3, 6, and 10 min at 700 300 microwave heating. In the latter case at 1 at
(C-3)
it
W W seen that is
sirnilar
goethite structure
servations. 3.4.
XRD
studies carried of the confirm phases determine and present. out to Microwave samples revealed the presence of hematite and magnetite and also disappearance of goethite. Additionally, these samples revealed stronger intensity lines of hematite, comparedwith raw ore, and weakening lines of magnetite. The ultimate product of heating iron oxides (in air) is hematite.6) The immediate product of goethite immediate after dehydration (about 300'C) grives a results
of raw ore and microwaved
samples.
Microwave time
Phases present
(min) Nil (*aw ore)
l : 700W
W W
2: 4:
700 700
6:
1300W 1300W
lO:
Hematite, Hematite, Hematlte, Hematite, Hematite, Hematite,
Magnetite, Goethitc, Magnetlte, Goethitc
Magllemite
Magnetite Magnetite Magnetite Magnetite
to that
observed after heating in air appears. This result can be explained by the fact that magnetite is a good microwave absorber and heated rapidly, whereas goethite is a poor microwave absorber which was mostly found between the magnetite grains and was heated gradually after the surroundingr grains were at hit)(rh temperature. It should be noted that the structural conversion of goethite to hematite was also confirmed by SEM ob-
XRDAnalysis Table 2 gives the
XRDanalysis
2.
grains
hematite-1ike pattern and on further heating reflection gradually becomessharper until at 900'C a typical hematite develops. The pattern of well-crystallised samples under of magneto-hematite heating curves 1, clearly demonin Fig. shown microwave power, as strate that dehydration temperature is reached immediately in the beginning of microwave exposure. It is apparent that the bound water is released from the oxides before the onset of further stages of microwave heating. However, for someoxides the contact
time with air may not be suflicient transform the phases to hematite. 3. 5.
to
completely
Reduction
Detailed results of
all
the reduction tests carried out
have been presented elsewhere7) and only is given here. For t,he samples pre-treated at both power levels used the degree of reduction creased with microwave treatment time, maximumand then decreased at longrer
a
summary
microwave at
first
in-
reached a treatment 13
ISIJ International,
times. In other words, the effect of microwave irradiation showecl a maximum at a certain microwave Specifically, maximum~g'ains over exposure time. those of untreated samples were and above
l)
Microwave pre-treatment 2min 5 olo Microwave pre-treatment
2)
6min
12
at
700 V\r
Vol.
31 (1991), No. l In general, both in the microwave samples as well as the
raw ore, flne porous metallic iron surrounded by dense metallic structure and often dense wustite particles were entrapped inside the iron shell. The
mayexplain why the
latter
at
olo'
observed7) rate of reduc-
becamemuchslower after 90 o/o ofreduction. However, for 10 min of 1300 microwave treatment, the difference becameobvious as shownin Fitg. tion
1300 W
W
The phenomenonof these beneflt maxima have been observed before in the reduction of Fe203CaC03composites,2) as well as in microwave heating and drying of t)crranular materials,8,9) but the exact reasons are unknown. However, the extent of this of the appears to be related to the permittivity materials involved. The results of the reduction of identical samples for various conditions are shown in Fig. 4 which exemplifles the principal features referred to above.
20
effect
I
Microscopic Observation Micrographs of reduced samples are shownin Fig, 5 for untreated ore (A), 700 (B), and 1300 (C) for microwave pre-treatment times of 2, 6 and 10 min (1-3) for the same reduction time in CO. Results of reduc.tion degree are given in Table 3, 3.5.
.
W
W
o o
o
e
15
o o e
~~~s)
,
io
',:
~ io
,
o
e
i:,
~
o
ieeo
e
o
5
•o
,
L,
o
•
OreA
A
A;
6mins.
MW
•
A; 10 mins.
MW
o
A, heating in air
cF
5
1O
15
20
Time, min
Reduction curves for Lampungore. crowave pretreatment power)
Fig.
(1
Wmi-
5.
of Lampungore after (See text for details)
Microstructure
14
300
reduction.
ISIJ International,
Table
Vol.
Reduction degree of microwave pre-treated samples at 700 and 1300 power. Reduction conditions : T=1 OOO'C; t=90 min ; 100 o/o COgas
3.
l
31 (1991), No. Table
Results of modified drop
4.
W
Pre*treatment time at 700 (min)
W
2 o 88
6 o 86
.
.
lO
o.84
Pre-treatment tlme at 1300 (min)
W
2 o 86
6
10
0.96
o 82
Sample
(A).
+6.3
-6.3
-3.35
mm mm - 500 mm +3 35 +500 um mm um .
Rawore
Rawore Heating
o 84
.
.
5(C-3) in that, Iess finely porous metallic iron appeared and a thicker iron shell around the wustite grains isolated them from direct contact with the reducing gas and hence hindered the reduction rate. This may explain why the overall reduction of this sample was lower than both the other two treated samples and the raw ore, Unfortunately, as with other unusual effects of microwave treatment reported in the literature,2,8,9) the reason why such a structure formed in one case and not in the others, cannot be
98 in air
96
I h)
(700'C,
treatment MW 300 W, 6min) treatment MW 300 W, 10min)
97 93
(1
Reduced* (raw Reduced (heating in
60
ore)
air,
700'C)
Reduced (6 min Reduced (lO min
MWtreatment) MWtreatment)
.
.
.
.
.
8 O
O,8
0.3
0,1
I,1
I .8
l,l
3
0,9
2 6
3.2
69 55
.
.
4,7 3.7
71 .2
l ,8 2.5
l,l
9.2
25
14.
l
5
4
6.4
9.6
14,6
3
10.0
17.2
17.5
1oo
25
Strength Measurements Reduction results presented above showthat microwave treatment increases the reducibility of the ores. However, before any potential application of this method of reducibility enhancementof lump ore feed to blast furnaces can be considered it is also essential to consider the effect of the treatment on the strength
.
ll.O
* Reducing atmosphere : 100 o/o CO, 1OOO'C Reduction degree : Constant (20 o/o)
explained.
5.2
Ore
Weight per cent
(1
3.
test.
.
Since the standard testro) requires greater amounts of ore than was available in the present study, it was decided to carry out a modified drop test using 100 g samples. It is considered that as each sample was subjected the results (Table 4) can be test, to an identical
_"'1a
20
go
"'1~Er'l'l~
--a--
15
Reducibility
-~-
~::
.~
80
Strength ~:~S)
of the particles.
comparedwith each other. If the standard level of strength of the ore
~?
~~:~s)
o C~
~) ':~
70
10
$:,a)
>
~1)
5
60
va].idly
is
taken
then all the unreduced samples are +6.3 as mm this be level-and below it after reducabove to seen tion. However, after reduction, air heated and microwave heated samples for 6min were stronger than the raw ore. If the heating treatments, both conventional and microwave, are viewed as a source of decrepitation
90
o/o
in subsequent reduction tests, then it interest of decrease in to note a signiflcant may be decrepitation (-500 um) by prior air and microwave
susceptibility
heating.
The reduction and strength indices in this case are combined in Fig. 6 from which it is obvious that the 6min microwave treatment is optimal. It
relate
should be noted that the results presented so far to the ore presently mined in Lampung. Be-
cause in future this ore (ore (A)) may have to be supplemented by that of a higTher goethite content5) from adjoining reserve, the microwave behaviour of this ore, here termed ore (B), was also investigated and the results are shownin Table 5 and in Fig. 7. Clearly, optimum is now seen to be at 10 min of microwave pre-treatment. The reason, both for the effect microwave on reduction and in the subsequent
50
o
o
2
4 Treatrnent, MW
Fig.
6.
Effect
6
8
Io
mirr
ofmicrowave treatment (1 300 W) on reducand strength. Ore A
tion time
strength of these two ores, is their different composition, as discussed by Standish and Huangll) and, especially with respect to goethite, by Watanabe and Yoshinaga,12) respectively. The formerll) investigated microwave application in carbothermic reduction of iron ores and noted that the differences in the observed reduction results were associated with differ* ences in the composition of the ores. The latterl2) studied abnormal behaviour of someore constituents Specifiand their effect on blast furnace operation. it cally, that goethite found dehydrated at as was about 350'C, the expulsion of water vapour in the associated with case where goethite was intimately that the ore the hematite aggregates, was so difficult in vapour decrepitate forced increase due to to an was However, when the ore was composedof pressure. large amounts of goethite the water vapour could be rather easily expelled without decrepitation.
15
ISIJ International,
Table
Results of modified drop
5.
test.
Ore
31 (1991), No. l
Vol.
25
(B)
1eo
Weight per cent
Sample
+6 3
-6.3
-3.35
20
500 mm 35 +mm 500 3 um mm mm um
90
s~
.
.
Rawore Heating
99 in air
I h)
(700'C,
treatment MW 300 W, 6min) treatment MW 300 W, 10 min) (1
(1
Reduced* (raw Reduced (heating in
.
99
ore)
air,
700'C)
Reduced (6 min Reduced (lO mln
MWtreatment) MWtreatment)
.
8 8
0.l
O.l
0,l
0.l
l,1
l.8
1, 15
80
'~ ~~~S) ~~~s)
97.5
l
.2
2 83 4 90 O
2.7
5,8
8,l
l
6,0
2.2
72.
1
4.2
8,
74
6
7.9
12,6
93
.
.
.
.
3.2
.8
2,5
1
o
C~ >
10
Cl)
5
- -
-t
15.6
4.9
a-
o o
Reducibility
of this study show that: Microwave treatrnent enhancesreduction and post-reduction strength, but beneflt optima exist. (2) In the present ore an optimumvalue of these
2
4
siderable effect
on the microwave treatment,
The oxidation
of the goethite present in the ore is obtained after the magneto-hematite grains are almost entirely oxidized to hematite. (5) Microwave treatment could be an alternative sintering to or pelletizing of the Lampungore to imits reducibility by suitably adjusting the treatprove time and ment power level of the irradiation. (4)
Io
of
J.
W. Walkiewicz.
S. L. McGill and L. A. Moyer: Symp. on " MicrowaveProcessing of Materials ". Mater. Res. Soc., Spring Meeting, April 1988, Reno, NV, Paper No. M4.7,
Mater. Res. Soc.
H. K. Worner and N. Standish:
3)
S. L.
McGill, J.
Analyst,
114 (1989),
115.
W. Walkiewicz and G. A. Smyres: Symp.
on " Microwave Processing of Materials ", Spring Meeting, 1988, Reno, NV, Paper No. M4.6, Mater. Res. Soc.
April
BHPTech. Bull., 25 (1981), Metallurgi, 5 (1985), 7.
4)
J.
Ostwald:
5)
lr.
Rustiadi:
6)
7)
G. Brown: Crystal Structure of Clay Minerals and Their X-R,ay Identification, Mineralogical Soc. MonographNo5, Mineralogical Soc., London, (1980), 6. Pramusanto: Ph.D. thesis to University of Wollongong,
8)
N. Standish, H. K. Worner and D. Y. Obuchowski: Elec-
(1989)
4.
.
I (1988), 45. Proc. Ist Australian Symp. on " Microwave PowerApplications ". Univ. Press, Wollongong, 1989), 227. J. K. Wright and A. L. Morrison : SEArSIQjuat., 14 (1985), tromagnetic Ener. Rev.,
Acknowledgement of a postgraduate study award for (P) of by the Australian International Deus one velopment Assistance Bureau is gratefully acknowledged.
50
8
microwave treatment (1 300 W) on reduction time and strength. Ore B Efrlect
7.
2)
9)
The provision
6
REFERENCES 1)
)
parameters was obtained with microwave treatment for 6min at a power level of 1300 W. (3) The presence of magnetite in the ore has con-
80
Strength
Treatment, min MW
Fig.
The results
lO)
N. Standish:
(
51.
ll) 12)
N. Standish and W. Huang: to be published in ISIJ Int., 31 (1991), No. 3. S. Watanabeand M. Yoshinaga: Trans. Soc. Min Engrs., 15 (19 68),
16
-
Conclusrons (1
~ ~::
O.2
* Reducing atmosphere : 100 o/'o CO, l OOO'C Reduction degrce : Constant (20 o/o)
4.
eJ)
70
3.