Post on 28-Jul-2020
\THE OXIDATION OF ILMENITEÜ A KINETIC STUDY
by
Lawrence J.“Corsa%{III
Thesis submitted to the Graduate Faculty of the
Virginia Polytechnic Institute and State University
in partial fulfillment of the requirements for the degree of
' MASTER OF SCTENCE
in
Materials Engineering
APPROVED:
‘/’ _/.
. / 0 /0 ¢ /T. P. Floridis, Chairman
‘ /
y.W7?J-L- Lyrrb . Hibbard
"7
. pence , pt. Head
June, 1977
Blacksburg. Virginia
ACKNOWLEDGMENTS
gc The author wishes to express his gratitude to Dr. T. P. Floridis
éof the Materials Engineering Department of VPI&SU. Dr. Floridis
Sinitiated this research and provided funds for my assistantship.
E;Without his technical advice and assistance this thesis would not
aéwhave been possible.
vsThe author also extends thanks to Drs. J. L. Lytton,
C. W. Spencer, and M. R. Louthan, Jr. for their wisdom in times
of disaster, and to the other graduate students in the department
for their help andfriendship.All
X—ray analyses were performed in the facilities of
Dr. C. R. Houska. The results obtained were necessary to the
research, and the author is grateful for the use of the equipment.
a Many other people contributed their patience, advice, and
assistance to the successful conclusion of my stay here at VPI&SU.
To all these people I sincerely express my thanks and friendship,
especially to J. T., who suffered more for this than I.
ii
1,
TABLE OF CONTENTS
Eää
ACKNOWLEDGMENTS ......................... ii
LIST OF FIGURES ......................... iv
LIST OF TABLES.......................... vi
INTRODUCTION........................... 1
REVIEW OF LITERATURE....................... 3
The System Fe—Ti—O ..................... 3Ilmenite and its Oxidation Products............. 5Kinetic Methods and the Oxidation of Ilmenite........ 10
EXPERIMENTAL PROCEDURES ..................... 12
Oxidation Study....................... 15I
Product Determination............. Q ...... 15Oxidation Kinetics ..................... 16
RESULTS AND DISCUSSION.............. Q ....... 19
_ The Oxidation Products of Ilmenitc ............. 19The Kinetics of Ilmenite Oxidation ............. 29
CONCLUSIONS ......................._ .... 65
RECOMMENDATIONS FOR FURTHER STUDY ................ 67
BIBLIOGRAPHY........................... 68
VITA............................... 7l
ABSTRACT4
iii
LIST OF FIGURES
Figure . Page
l. The system FeO—Fe203-Ti02 showing the three solidsolution series. At low temperatures the joinTiO -Fe O becomes stable................. 42 3 4
2. The system FeO—TiO2 at oxygen.pressures defined bythe iron-wüstite equilibrium ............... 6
3. The system Fe-Ti—0 showing Fe 0 -Ti0 and Fe0-Ti0. . . 2 3 2 2binary joins ....................... 7
4. The equipment used in the thermogravimetric analysis . . . 17
Sa. Ilmenite as synthesized (-325 mesh) 500X ......... 25
5b. Ilmenite as synthesized (-325 mesh) 2000X......... 26
Sc. Ilmenite oxidized 24 hours at 520OC (-325 mesh)...... 27
5d. Ilmenite oxidized 24 hours at 790OC (-325 mesh)...... 28
6. The percentage of theoretical oxidation versus time forthe 45-75 um particles .................. 46
7. The percentage of theoretical oxidation versus time for75-150 um particles.................... 47
8. The percentage of theoretical oxidation versus time for150-180 um particles ................... 48
9. A plot of temperature versus time for the kineticsamples.......................... 51
10. The percentage of theoretical oxidation versus time forall three particle fractions as detcrmined from a'piecewise‘ least squares technique............ 54
ll. A plot of oxidation rate (log d(% ox)/dt) versus. 10reciprocal absolute temperature.............. 57
12. A plot of oxidation rate (normalized to interfacialarea) versus reciprocal absolute temperature ....... 58
iv
Figure Page
13. A plot of oxidation rate (normalized to product layerthickness squared) versus reciprocal absolutetemperature........................ 59
v
LIST OF TABLES
I. Analysis of Reagents ................... 13
II. The Interplanar Spacings Obtained Experimentally From-325 Mesh Synthetic Ilmenite ............... 20
2III. The Interplanar Spacings of Ilmenite and Its Reported
Oxidation Products .................... 2l
IV. Experimental Results for Samples ............. 30
V. Weight Gained....................... 50
VI. Calculated Results for the Oxidation of Particles..... 60
VII. Calculated Resulted Used in Determination of ActivationEnergies by the Method of Linear Regression........ 64
vi
INTRODUCTION
Titanium is rapidly finding use as a non—aviation construction
material largely on the merits of its excellent mechanical properties.
No metal since aluminum has found such a wide variety of applications
so soon after its commercial introduction. Unlike aluminum, no
economical process for the industrial manufacture of titanium metal
has been discovered. This prevents titanium from being used in many
products which could benefit from its use and keeps it in the category
of an exotic material. For these reasons, a cheaper, more efficient
method of winning titanium from its natural state is of large concern _
to the industry. °
The two minerals most commonly processed for their titanium
content are rutile (Ti02) and ilmenite (FeTiO3). Of these, ilmenite
is by far the cheapest and most abundant material(l) but is less
desirable from the standpoint of processing. To date, industry has
used ilmenite mostly for the production of pigment grade rutile,
but as the reserves of rutile ores have become less readily available,
the gradual substitution of ilmenite ores has increased(2).
The major processes currently used for processing ilmenite are(l):
(a) Smelting with coal or coke in an electric furnace to produce
molten iron and a titanium—rich slag (the Sorel slag process)
(b) A selective chlorination using hydrochloric acid or a
combination of chlorine and carbon monoxide gases at high
temperatures and pressures _
l
2
(c) Sulphidization using either sulfuric acid, hydrogen sulfide
with carbon, or sulfur vapor at high temperatures and
pressures. l(d) Solid state reduction of the iron oxide component in the
presence of a catalyst (the Becher process).
These processes end with a concentrate high in TiO2, which is then
processed for titanium metal. Only the Becher and Sorel slag methods
have been used on an industrial scale.
The direct treatment of ilmenite for the extraction of titanium
is considered desirable, but an economical process has not been
developed. Direct reduction processes are quite sluggish(3’4),
although prior oxidation treatment holds much promise for the
In order to optimize this type of process, a thorough
study of the oxidation process in ilmenite is of primary importance.
The thrust of the present investigation will be to determine the
kinetics of the oxidation process and the end products of roasting.
REVIEW OF LITERATURE
Previous work pertinent to this investigation falls into three
basic categories:
(l) Studies of the equilibrium system iron—titanium-oxygen and
its subsystems
(2) Studies of the mineral ilmenite and its alteration products
(3) Analytical methods for the determination of kinetic data
using non-isothermal techniques.
The System Fe—TieQ
Much of the thermodynamic and equilibrium data available in the
literature were determined for the pseudo-ternary system 'FeO'-Fe203—
Ti02 at temperatures above 9OOOC(7’8’9). This system has beenshown(lO)
_ to have three solid solution series in this range of temperatures
(see Figure l):
(l) The cubic system magnetite—ulvospinel (Fe304-Fe2Ti04) series
(2) The rhombohedral hematite—ilmenite (Fe2O3-FeTiO3) series
(3) The spinel pseudobrookite (Fe2TiO5-FeTi2O5) series.
A fourth solid solution series, magnetite—rutile (Fe3O4—TiO2) is found
at temperatures below aoo°c(ll). The pseudobinary join Fe2TiO5—FeTi2O5
was first shown to be a complete solid solution series above ll5OOC
by Akimoto, Nagata, and Katsura(l2) and has since been confirmed(7’l3).
FeTi2O5 is an unnamed compound not found in nature. The join Fe203—
FeTiO3 exhibits solid solution above 9500C, and the join Fe304—Fe2TiO4
is a complete solid solution series at temperatures as low as 6OO0C(l4).
3 p
A
TiO2
FeO'2TiO§(FeTi2O5
FeO·TiO(FeTiO3§
‘ Fe O ·TiO
FeO Fe O2 3(Fe304
Figure l. The system FeO—Fe 0 —TiO showing the three solid. . 2 Q 2 , .- solution series. A low temperatures the %oinTiO2—Fe304 becomes stable. (From Lindsley 11))
5
The join Fe0—TiO2 (Figure 2) of the Fe-Ti-O system (Figure 3)
is of major importance to this study since ilmenite is but one of
three compounds in this binary system. Each of these compounds is
an end member of a solid solution series in the Fe2O3—'FeO'—TiO2
system. An early study(l5) revealed only the compounds ulvospinel
(2Fe0·TiO2) and ilmenite (FeO·TiO2), but more recent research(l2’l6)
demonstrated the existence of FeO·2TiO2.
Ilmenite and its Oxidation Products
There is much apparent disagreement in the literature concerning
the behavior of ilmenite at temperatures below lO0OOC, as pointed out
by Rao and Rigaud(l7). Much of this is traceable to a fundamental
difference in the samples used. Experiments involving the oxidation
products of ilmenite are performed for either mineralogical or
metallurgical reasons, with the former preferring to use naturally
obtained ilmenite concentrates. Temple(l8) has investigated several
of the commercially available ilmenite ores and found that they were '
not ilmenite ('FeO' content 48%), but 'altered' ilmenite, or leucoxene,
with 'FeO' contents ranging from 0.l% to 49%. He speculated that this
alteration took place by slowly progressing oxidation and leaching of
the iron compounds. In any case, it cannot be expected that these
materials would behave like chemically pure FeTiO3.
Recent investigations involving synthetic or hand—picked grains
of natural ilmenite report the following groups of reactions:
6
. Pseudobrookite+
Liquid' tI°l1500 g[ }“+‘ E
^ Il '1 1* *6 Iow Ulvospinel 1‘“+W ' L' 'dusiite Liquid)
I Rutilg \Q Liquid Ä IE3 J ' \ ' 1Q \ Ulvo— IhlmenitqIQ
[ . I I brookiteH spinel [I + [ [ +
Il3OO wustite [ : + I?seudo—1 [ R tile
1 _ [ glm€nj;qbro0kitd1 U1 Wustite [ [
J TI + 1 1 .Il -IUIvospinel I
§hVOSpin2inltC Pseudobrookite
FeO 20 40 60 80 TiO2Z TiO2 +
Figurs 2· The system Fe0—Ti02 at oxygen pressures definedby the iron—wüstite equilibrium. (From MacChesneyand Muan(l6))
AA AVAAVAVAAVAYAYA
AYAVAYAYA
AVAVAHVAVAVAYAAVAYAYAVAVAVAVAYA
Te AYAYAVAYAYAYAYAYAYA T
8
. 17 19(1) as suggested by Rao and R1gaud(
’ )
(a) below 770OC, hematite and rutile form
4FeTi03 + O2 + 2Fe203 + 4TiO2
(b) from 7700C to 8900C, pseudorutile and hematite form
12FeTi03 + 302 + 4Fe2Ti3O9 + 2Fe2O3
(c) above 8900C, pseudobrookite and rutile form
4FeT103 + 02 + 2Fe2T105 ZT1 2
(2) as suggested by Grey and Reid(20), Ramdohr(2l), Curnow and
Parry(22), Teufer and Temple(23), and Overholt, Vaux, and
Rodda(24)
(a) below 800OC, pseudorutile and hematite form -
(b) above 800OC, pseudobrookite and rutile form
. (25)(3) as suggested by Karkhanavala and Momin l
(a) at 650OC, pseudobrookite and rutile, 'pr0bab1y' with
small amounts of hematite
(b) at 850OC, a 1:5:7 molar mixture of hematite, pseudo-
brookite and rutile
l2FeTi03 + 302 + Fe203 + 5Fe2Ti05 + 7TiO2
Teufer and Temple(23) further demonstrated that pseudorutile decomposes
above 80OOC to form pseudobrookite and rutile by the reaction:
Fe2Ti309 + Fe2Ti05 + 2Ti02
and demonstrated the orientational relationships of an oxidized single
crystal to be
Ilmenite Pseudorntile Rutile
{0001}[11ä0]|1{0001}[10l0]||{010}[0011 .
9
The mineral pseudorutile has never been synthesized from its
constituent compounds, hematite and ruti1e(l9). Larrett and Spencer
(see ref. 17, p. 325) report that pseudorutile only occurred when
synthetic ilmenite was oxidized, but all agree that pseudorutile is
a phase distinct from both the dubious Fe203·3TiO2 compound 'Arizonite'
discovered by Palmer(26) and the mixture reported by Karkhanavala and
Momin(25). (It was pointed out by Temp1c(l8) that the unidentifiable
peaks reported by Karkhanavala and Momin obtained from the oxidation
products at 650OC and 850OC are clearly due to the formation of
pseudorutile, a compound not identified in 1959.) The inability to
synthesize pseudorutile directly is believed due to either the meta-l
stable nature of the compound and the corresponding need for a specific
reaction path with divalent iron as the diffusingspecie(2O)
or to
g the slowness of the diffusion process in the temperature range where
pseudorutile exists(l7).
The compound Fe2Ti05 was shown by Pauling (see ref. 12, p. 38
and ref. 20, p. 188) to have the orthorhombic pseudobrookite structure.
lt is an unstable compound with a lower temperature limit between 800
and 900OC. Rao and Rigaud(17) determined pseudobrookite to be unstable
with respect to pseudorutile at temperatures below about 900OC using
differential thermal and gravimetric analysis, whileothers(2O—24)
have reported the presence of pseudobrookite as low as 800oC by
actual experiment.
l0
The only investigations of the kinetics of the oxidation of .
ilmenite reported in the literature are those of Rao and Rigaud(l9)
and Karkhanavala and Momin(25), with the results of the former being
central to this thesis. The latter investigation amounted to a
single published thermogram obtained under conditions of linear
temperature increase. The principle features of this thermogram
were:
(l) small amounts of oxidation at temperatures below 30OOC
(2) a sharp increase in the reaction rate above SOOOC
(3) a failure of the reaction to reach its theoretical limit of
5.2% weight gain .
There was no 'abrupt weight gain above 8O0OC' as réporced by Rao and
. Rigaud in their study. Karkhanavala and Momin maintained temperature
at 850OC to allow the reaction to go to completion. This appeared
on their chart of weight gain versus temperature as a vertical line,
which was apparently mistaken for a sharp weight increase.
Rao and Rigaud were more thorough in their experimental work.
Using both isothermal and non—isothermal techniques, they determined
that ilmenite powders oxidized according to the logarithmic rate law
with an activation energy of 9.5 kcal/mole in dry oxygen. They also
determined that ilmenite platelets oxidized parabolically with an
activation energy of 20 kcal/mole up to 800OC, and 43 kcal/mole above
60006. The morphology of the product layer was studied using both a
scanning electron microscope equipped with an energy—dispersivel
ll
analysis unit for determining the distribution of elements and
platinum markers imbedded in the platelets. They report that oxygen
diffusion through the product layer is the rate limiting step at
temperatures lower than 900OC.
For this investigation, non—isothermal techniques were employed
in order to obtain large amounts of information from a small number
of samples. A short discussion of the technique is given by
. . (26) (27) .Kubaschewski and Hopkins and by Wood , both of whom mention that
the temperature must increase linearly with time. This condition was
required to make a mathematic variable transformation, which was not
shown to be necessary in the general treatment of Freeman and
(28) . .· . . Q .Carroll . Using thermogravimetric analysis even without linear
heating rates, good agreement with isothermal data has been
(29-31) . . . . .observed . A major problem with this method is that the reaction
mechanism must be determined by trial and error, with the rate law
determined from the most appropriate model(j2’j3).’
EXPERIMENTAL PROCEDURES
The techniques and equipment used in the course of this study
are presented in the following sections. All additional information
necessary to the reproduction of these data are also included.
Sample Preparation
The ilmenite (FeTiO3) required for these experiments was
synthesized using commercially available reagents using the following
procedure:
(a) Iron oxide (Fe2O3), as hematite powder, and titanium dioxide
(TiO2), as anatase powder (see Table I), were dried for
several days at l25OC to reduce the inaccuracies that could
be caused by moisture in the reagents.
(b) The powders were weighed on the Sartorius type 2842d
analytical balance which was capable of one-tenth of a
milligram precision. Twenty—five grams of mixture were
prepared in the exact stoichiometric proportions for the
reaction
2Fe203 + 4TiO2 = 4FeTiO3 + O2
(c) The powders were blcnded overnight under acetone and air-
dried. The mixture was ground with a mullite mortar and
pestle to achieve a finer, more homogeneous mixture.
(d) The material was packed into a four-inch quartz combustion
boat and placed in the uniform temperature zone of a Lindberg
'Hevi—Duty' furnace. The furnace was equipped with a mullite
‘l2
13
TABLE T, Analysis of Reagents
Titanium Dioxide (Ti02, as anhydrous anatase) 99.95% pure obtained
from J. T. Baker Chemical Co.
Water Soluble Saltsl
0.03%
Arsenic (As) 0.0001%
Iron (Fe) 0.002%
Lead (Pb) 0.006%
Zinc (Zn) 0.009%
Ferric Oxide (Fe203, as red, anhydrous hematite) 99.81% pure obtained
from the Fisher Scientific Company.
Arsenic (As) 0.002%
Nitrate (NO3) 0.01%l
Phosphate (PO4) 0.02%
Sulfate (S04) 0.05%
Manganese (Mn) 0.05%
Copper (Cu) 0.009%
Substances not p'ptd byNHQOH 0.04%
Zinc (Zn) 0.007%
14 ·
furnace tube and a Lindberg controller. The temperature
was measured with a chromel—a1umel thermocouple and a
potentiometer. The system was flushed with carbon dioxide
gas flowing at 300 cc/min for at least one hour prior to
the start of the run. The flow rate was then decreased to
180 cc/min. As the temperature increased, a 90% argon—lO%
hydrogen mixture was added to the gas stream. The gas
mixture was chosen so that, at 1000OC, the following reactions
would be occurring:
(1) the reduction of hematite to "wüstite"(34)
Fe203 + CO + 2'FeO' + CO2
_ Fe203 + H2 + 2'FeO' + H20
(2) the "water—gas" equilibrium(35): .
CO2 + H2 = CO + H20l
Titanium dioxide does not participate in the solid—fluid
equilibrium. Six days were allowed for the solid state
reaction
'FeO' + TiO2 + FeTiO3
to occur.
(e) The furnace was opened and the sample quickly removed at
24 hour intervals for the first four days. The sintered
material was broken and crushed with a steel impact mortar
and replaced in the furnace. This procedure prevented the
sintering process from hindering the gas infiltration, and
further reduced the chance of inhomogeneity in the material.
15
The material was left undisturbed for the final 48 hours
to allow sintering to consolidate the fine powders.
(f) The ilmenite was sampled and examined with X—rays to assure
that the desired reaction had gone to completion. Finally,
the mineral was crushed and sieved to obtain three particle
size fractions. The finest particles (-325 mesh) were used
in the X—ray identifications.
Oxidation Study A
The experiment was separated into two areas:
(a) the determination of the oxidation products and morphology
(b) the study of the rates of oxidation. ·
Product Determlnation’
Five-gram samples of synthetic ilmenite were returned to the
same Lindberg furnace used in the synthesis of the material. The
specimens were held under flowing dry oxygen at temperatures in the
range 50OOC — 850OC for periods of 12 to 24 hours. The oxidized
samples were air—quenched and subjected to an X—ray diffraction
analysis. The Siemens diffractometer was equipped with a graphite
crystal monochromator to filter the KB component of the copper K
radiation. The sample was held in a device which tumbled and
oscillated the loose powder specimens to prevent any effects of
preferred powder orientation. The scans were performed at or 1
degree/minute under 50 kV radiation. A scintillation detector was
16
used for maximum sensitivity to the weak diffracted intensities
produced.
Oxidation Kinetics
Powdered ilmenite specimens weighing from 300 to 400 mg were
loosely placed in a quartz crucible.‘ A range of sample weights was
used in place of a single weight to determine if the height of
material in the crucible had an effect on the course of oxidation.
For example, if the particles were sintering as the temperature was
increasing, the sample height might eventually affect the path length
for the penetration of oxygen to the bottom of the crucible. „
An Ainsworth Recording Balance system, type Rv—AU-2, continuously
monitored the weight and temperatures of the specimen. The system
employed a Lindberg 'Hevi—Duty' furnace mounted vertically beneath4
the left pan of the balance (see Figure 4). The final temperature of
the run was preset on a Honeywell type 54261 controller whose thermo-
couple was placed in the furnace insulation at the level of the hot
zone.
The sample was placed in a platinum basket which was suspended
from the left pan by a platinum wire. The temperature of the specimen
was monitored by a sheathed chr0mel—alumel thermocouple located 7 mm
below the sample, well within the 4 cm uniform temperature zone of
the furnace.
The entire system, including the balance enclosure, was flushed
with dry oxygen for at least thirty minutes before each run. The p
17
transducertares
1-—upper gas outlet
gasinletquartz
furnace tube.?
furnace-:2
specimen holder
thermocouple sheath
V lower gas outlet -j2;,l
. ‘;S- rhermocouple leads
Figure4_ The equipment used in the thermogravimetric analysis.
18
electronics were allowed to equilibrate during this time. The flow
rate was reduced to 200 cc/min and the balance enclosure gas outlet
was closed at least two minutes prior to the application of power to
the furnace windings. No change in sample weight was detected during
this period.
The experiment commenced when power was applied to the furnace
winding. A Variac, rather than a switch, was used to prevent sudden
changes in line voltage which could affect the equipment. The high1
inertia furnace was allowed to heat at its own rate, which varied
continuously from 3500/min to less than 10OC/min as the experiment
progressed. Data were taken at 25OC intervals from 30OOC to 8500C,
at which time the experiment was terminated. Each run required 50 to
55 minutes from start to finish, depending upon ambient temperature.
This procedure was repeated for five runs each of three particle
lsize fractions. Four standardizing runs were also made to compensate
for convection current and bouyancy effects encountered as the gas
passed down through the hot zone of the furnace. Titanium dioxide,
which had been dried and vacuum degassed, was chosen as the standard
material since it is inert under oxidizing conditions over the
temperature range of interest. The weight of all standards was
checked with the analytical balance before and after each run with
no change noted. The weight increase of each sample was also checked
and only small discrepancies were noted. These could be due to
additional oxidation taking place during the six hour period required
to cool the furnace. 4
RESULTS AND DISCUSSION
The results of this study are best presented in two parts for
the purpose of discussion:
(l) the determination of the oxidation products
(2) the determination of kinetic data for the oxidation reaction.
The results will be correlated in the next section.
The Oxidation Products of Ilmenite
X—ray diffraction analysis was performed as previously described
to determine the phases present at the various temperatures. The
results of this work are collected in Table II. Table III is provided
for purposes of comparison and contains the interplanar spacings of
the reported oxidation products obtained by some previous investiga-
tions. It must be remembered that the diffracted intensity from
these specimens is quite small due to the very fine—grained nature of
the oxidation products. This tends to make the material very amorphous
in character, causing the X—ray peaks to generally be broad and
diffuse(l8’36). As a consequence of this, no accuracy better than
i_0.02X is claimed for the experimentally determined lattice spacings.
No intensity data are reported because no reliable data could be
obtained from the X—ray charts.
The synthetic ilmenite had no discernable peaks due to the
presence of impurities, unreacted material, or other iron-titanium
compounds. Magnetite and ulvospinel have occasionally been found
by other investigators(l7) and in previous synthesis attempts by
. l9
20
TABLETI_ The Interplanar Spacings Obtained Experimentally From
-325 Mesh Synthetic Ilmenite.
Sample Sample Sample Sample Sample Sample1L3—1 IL3—2 IL3—3 IL3—4 1L3—5 IL3—6as syn— oxidized oxidized oxidized oxidized oxidizedthesized 24 hgurs 24 hours — 24 hgurs 24 hgurs 12 hgurs
520 C 615OC 710 C 790 C 850 C
2*89 2*89 2.88 2.88 2 862.74 2.74 2.73 2.74 2.74 2.732.70 2.70 2.70
2.54 2 532.68
° 2.52 2.52 2.52” 2.50 2.502.47
2.43 2°99 2.43 2.43 2°992*29 2.23 2 212.20 2.20 2.20 2°l9' 2.11 '2*29 1.961.87 1.87 l 851.84 1.84 1.84 l°741.73 1.73 °1.71 1.70 1.70 1.71
1.69 . 1.691.67 1.67 l 661.64 1.64 1.64 1.64 1.64 l°631.60 2*2;
1.50 1.50 g1.50*
1 47 1 47 1.48 2*99 1.48 9841.46)° ° 1.45 1.45 1.45 DB 1.451.43 1.43 1.431.37 1.37 1.37 D8/2*gä
1.34 1.34 DB=1.34>
*Diffuse Band
21
TABLE [[[_ The Interplanar Spacings of Ilmenite and Its ReportedOxidation Products.
Hematite Ilmenite Rutile Anatase Pseudobrookite Pseudorutile(Fe209) (FeT1O3) (T1O2) (T1O2) (Fe2T1O5) (Fe2T1309)
ASTM No. ASTM No. ASTM No. ASTM No. ASTM No. From Gääö &13-0534 3-0781 21-1276 4-0477 9-0182 Reid 2l 2.88
2.74 2*75 2.742.69
2*5“ 2.502.51 2 49° 2.452.43 2.432.38
‘ 2. 42 39
3 2.312.23 ° 2.22
2.20 2.19 2.20~
2.202.12
2.07 2.95 1 2.011 89 1.97
1.86 ° 1.861.84 1.741.72
1 79 1.711.69 1.69 ° 1 671.66 1.66 _ l'6&1.63 1.63 1.63 1.63 °
1.621.60 1.54
1 591.51
° 1.49 1.49 1.491.48 1.48 1.481.45 l'“7 1.45 1 431.42 1.42 °
1.35 1.36 1.37 1.371.34
1.31 1.31
22
the author. A crude chemical analysis was accomplished using energy
dispersive analysis while the specimen was being examined in the scan-
ning electron microscope. Roughly equal amounts of elemental iron
and titanium were present at all points along the several grains
tested.
The lines obtained from the sample oxidized for 24 hours at
520OC are still predominantly those of ilmenite, but some new lines
have appeared with "d" spacings of 2.20X, 2.43X, and 2.89X. The
line at 2.432 could be attributed to anatase, and the 2.20X line
could be hematite, but all three taken together indicate the presence
of pseudorutile. The absence of lines of 2.12X, 2.012, 1.70X and
1.672 makes positive identification complicated, but no other
reported iron-titanium compound has a lattice reflection at d =
2.89X.
The sample oxidized at 61500 for one day displayed some lines due
to hematite, and some due to the presence of either unreacted ilmenite
or increased amounts of pseudorutile. As before, there are lines
missing from the pattern of pseudorutile. but there are also lines
missing from the patterns of the other possible mineral patterns.
At 710OC, the pattern obtained from the oxidation products is
certainly that of pseudorutile and hematite. The line at 2.89X has
been gaining steadily in intensity as the temperature was increased,
and the lines have become sharper, particularly the low angle lines
(large "d" spacing).
23
The lines due to pseudorutile and hematite are present in the
X—ray pattern of the specimen oxidized at 790OC, and generally all
the lines present have become sharper and more intense. The dis-
appearance of the lines at 2.50X and 1.60X are exceptions to this
rule, as is the reappearance of the (019) pseudorutile reflection.
The sample examined at 850OC was reacted only twelve hours as
Karkhanavala and Momin(24) had reported that no changes occurred in
their diffraction patterns after six hours at this temperature.
However, the pattern obtained by this investigator was considerably
more amorphous in character than the other patterns. Few of the lines
were either intense or sharp. Several diffuse intensity bands were
noted. These can be attributed to either several small peaks over-
lapping or to the presence of solid solution. This discrepancy could
be due in part to the increased thickness of the plastic material used
to contain the loose powder while the X-ray analysis proceeded. The
accuracy obtained from this determination is correspondingly less
than from those patterns taken at lower temperatures. However, the
presence or absence of intensity was still detectable. Pseudorutile
was present in this sample, as shown by tie low angle lines, but the
phases pseudobrookite and rutile also appeared at this temperature.
Clearly all three phases cannot coexist in equilibrium, but there is
no reason to expect that equilibrium was reached by the reaction.
In summary, the following has been demonstrated:
(a) at temperatures as low as 520OC, the mineral pseudorutile
may be present. The reaction is probably quite sluggish ·
ZA
at this low temperature, indicating that perhaps as much
as a month may be required to attain equilibrium.
(b) at 6l5OC, hematite is clearly present. Lines due to either
ilmenite and rutile or, more probably, to pseudorutile are
quite sharp.
(c) at 710OC and 790OC pseudorutile and hematite are the only
phases present.
(d) at 850OC, pseudorutile is present to some degree, but hema-
tite is no longer detected. Instead, pseudobrookite and
rutile have appeared.
The-results of the X—ray study indicate general agreement with -
reported phases. The following conclusions may be drawn:
(1) no lower temperature limit exists at 77OOC for the formation
of pseudorutile, as reported by Rao and Rigaud(l7). Bothl
this investigation and those of Temple(l8) and Karkhanavala
and Momin(25) show the occurrence of this mineral at temper-
atures well below 800OC.
(2) Pseudobrookite occurs at temperatures below 890OC. This
again conflicts with Rao and Rigaud, but agrees with Grey
and Reid(2O) who reported it at 800OC.
The morphology of the ilmenite and its oxidation products can be
seen in Figures 5a—d. It must be noted that the microscopy was
performed on -325 mesh particles, including fragments and other fines.
The fines adhere to the particles making them appear more jagged and
25
Figur~ Sa. Ilmenite as synthesized (-325 mesh) SOOX.
26
Figure Sb. Ilmenite as synthesized (-325 mesh) 2000X.
27
Figure 5c. Ilmenite oxidized 24 hours at 520°C (-325 mesh) 2000X.
28
Figure Sd. Ilmenite oxidized 24 hours at 790°C (-325 mesh) 2000X.
29
broken than they actually are. They would also hide the appearance
of a fragmented oxide layer, if one exists.
As synthesized, the particles are generally rounded in shape with
an uneven surface. No cracks in the surface are apparent even at
higher magnification. After oxidation at 520OC, the surface features
are "softer" in appearance and there are no signs of surface fracture.
llmenite oxidized for 24 hours at 79OOC did show signs of high
stresses. The surface appears mottled and jagged, and cracking is
_ apparent. The product morphology is important in a study of the
kinetic behavior of any material(l7) as it may determine the mechanism
that controls reaction rate.
The Kinetics of llmenite Oxidation
The results of the thermogravimetric analysis of each sample are
given in Tables 4a-o and are shown graphically in Figures 6a—c.
Although there is some scatter in the data, a definite trend is
clearly shown. The scatter is due to improper standardization of
the equipment. This source of error can be explained by describing
the results of the standardization runs.
Each titanium dioxide standard was performed as if it were an
experimental run. The same experimental procedure was followed for
all runs, but the standards did not give repeatable results over
the initial stage of the experiment. Their apparent weight gain
always peaked shortly after power was applied to the furnace, passed
through zero and then stabilized at some value. All this occurred
30
NI
Or—II
Oou 1In -,-+:1 ~7'®¤’\OC‘<."f'*·U‘~7'bx7'f\IO<\1C~~1'€\I<¥:«—+·—4fI)v—¢O
••••••••••••••·-••••••••,.., ,_,.,.,¤, N~Mmm¤‘•m~1·¢~1·+r•~¤+~¤~=">«wc¤m<rmm+:x0~
pqpqg., »~4r·I¤-·¢(\Jt'(‘•¢"lÜ~OI‘MOV
rxFU <r
U)4-IN
00*:* ••••••••••~•••·=••••••••
ß°EjéI>‘
¢»·rIn~0~¤—ot~+~«>c~cr~m~0i«”·mmr~~0co-•+~„3 E ,-+•-+~I•-·+„-+K\I¢‘I\~'?‘~O®C„*<\ß„<1—
•-I
3 E9 "v—I5 .U)aßM rx
’
aß-a"I gv f~•··I·O•-*!‘(*‘+¢Ä?"I‘·L¢\t<1•·~:•-·+f\I<"7+~OC‘€·‘>0‘I~¤~1'f"§~O<!'1 E,V
r··Ir·-4•-I•-»•,-4•-d«-:,-•«—4l‘\l(\¢\I¢”\l!\J¢\]l‘I’N3E
CUpc?
Q,,\ OL0Ou’\C)Lf\OU\OL!\OIf\"D'OOlI'\O'1\OL’\C‘LF\¢Ü\M EU ONl¤|*ONl!\|“•O€\JL¤Y~€7<\ßl{\f*CJNI!\h~C*<\!LV\
gov fGN'\¢¤M~T~1"~1'~$'II\U°\If\L!\~O~O~O~Of~I‘|*I~<ÜtD®
-3>+-4
M-1III .<CE-I
Nl
N‘T‘'SETJ‘*,:,1*5 :~··»~o~~¤~~m~\+c~«wc»¤m—:·»r~m<·-«o««¤m~
gl Q)-U}.;g•|•••••••••••<••••••••
Hggf; »-•-4,-:¤1m:-1-:·¤'x—0¤o
:-4CLE 6
ä U ‘*O~OU'\(\I>•-4<‘?°$~OO‘~OL!'\®ß’\(\j•··4r¤U°\O‘€T)u'\f*‘„f\L’\I~·
O „.(m><I•lI•|l•I||I•I0•I••••••
U-4 QJLDEV)
**0 •-•«—•g xx:-4IJ 4U}
rx ..
H O~1"U*ON-I|~N®N•-·4t-U\L¤C>ONO0'>O|*¤DU\
u Hm „g»0:~r~c0¤oO‘O«-<NNo"•¢~0r~¤OO~•v¤mb—¤#-«0:.3 V •·-e•-4·—••-a-«„«-n»:•-«¢x!n‘~J¤1<\1rx1¢"‘:«·¤w
‘
E-:-4;.; .
gf Ugi OLhOLOO'-GOUNOU\OlOOlhOL^O4~^®U\®dNÖ
moHVMf‘0Mr¤<4‘«1'~1'~1'L¤Lt\Lt\kI\~O~3~O~04‘4‘·P—I~'J9¤D€X3 »
‘ _
'
I—4•• I-rl
:-7gg .< .H
1
S?
'9!
T'T
?T
8ST
EC
S8
ST
'?9
?'9
IIV
TGZ
SZ
?S
E'Z
S9
'96
.E
69
ZC
CG
I!'Z?
8'8
!9
8?
ZS
L!
!Z°?
€Z
'E9
?0EZ
C9
!·
E9
°9Z
T'6
?E
?TZ
SZ
!9
!'CZ
E°8
EZ
66TC
C!
.?
?‘¢
TG
'CC
ES
SI
S!9
?§
°ZI
i‘EZ
9I!I
C9?
66
'6?
'8I
96ST
SZ
9l
9E
'8?
'9T
98
?T
00
9E
Z'!
E'E
ISG
ET9
!SQ
6Z
'99
'IIZG
ZTC
SS!9
'SS
°0I
06TT
SZ
9O
Z'S
9'6
?IT
IC
OS
~G
6'?
O'6
SE
OI
SL
?L
S'?
?'R
09
6O
9??
E°?
O'8
96
8S
Z?
_Z
T'?
9°!
SE
8C
O?
_?6'E
E'!
L9!SLE
6!'E
O'!
SE
!C
SE
89
°EZ
'!6
99
SZ
E‘
4!Z
'E_
C'9
9?
9O
CE
f
(quaoxag)(V
o;
xm
ä)(·0
ag
)(3
0)
uo
yqe
pyxg
up
eg
SM
IL·d
ma
l4
go1ua1xg
1qB;aM
4
{—gZ—
yyq
]a
jdm
eg
10
;sq
ynsa
gIe
qu
am
yxad
xg*0
A]
gq
qvl
TAB
LE1V
d.E
xp
erim
en
tal
Re
su
lts
for
Sam
ple
IL4
2-1
0-1
We
igh
tE
xte
nt
of
Tgmp
.Ti
me
Gai
n4
Oxi
datio
n(
C)
(Se
c.)
(gm
x10
)(P
erc
en
t)
30
06
72
5.3
3.3
7‘
I3
25
71
55
.93
.71
35
076
26
.64
.17
37
58
13
7.3
4.5
94
00
86
77
.84
.93
42
59
24
8.2
5.1
84
50
98
98
.55
.39
47
51
05
79
.15
.77
50
01
13
39
.45
.94
·5
25
12
15
10
.76
.74
550
1304
11.7
7.39
M5
75
14
01
12
.88
.11
w6
00
15
03
14
.79
.31
·6
25
16
15
17
.41
1.0
2.
„6
50
17
35
21
.61
3.6
76
75
18
69
28
.91
8.2
4.
70
02
01
14
0.1
25
.34
~·
72
52
15
95
2.6
33
.26
75
02
32
16
5.2
41
.22
77
52
50
77
9.3
50
.11
80
02
70
79
5.0
60
.03
825
2925
'1
13
.37
1.5
7_
85
03
17
01
32
.58
3.7
0
TAB
LE1V
e,E
xp
erim
en
tal
Re
su
lts
for
Sa
mp
leIL
42
-11
-1"
We
igh
tE
xte
nt
of
Tgmp
.Ti
me
Gai
n4
Oxi
datio
n(
C)
(Se
c.)
(mx
10)
(Pe
rce
nt)
30
06
86
7.0
3.3
43
25
74
58
.03
.79
35
07
77
8.7
4.1
43
75
82
99
.54
.50
·4
00
88
71
0.0
4.7
5·
42
59
51
10
.24
.85
45
01
01
41
0.5
5.0
14
75
10
83
11
.45
.44
”5
00
11
65
12
.15
.75
.52
51
24
51
2.7
6.0
2U)
55
01
33
91
3.8
6.5
65
575
14
41
15
.77
.48
·6
00
15
50
19
.59
.29
.6
25
16
62
23
.41
1.1
46
50
17
39
28
.81
3.7
16
75
19
32
40
.81
9.3
870
02
07
15
3.8
25
.58
725
22
23
70
.43
3.4
87
50
23
93
88
.34
1.9
9'
775
25
83
_1
08
.25
1.4
48
00
27
91
13
2.4
62
.94
82
53
02
51
58
.27
5.1
985
0p
3299
18
3.1
87
.03
E —r'FJ
1-4,<*‘*‘·A
I/\!—]FJOmbq
OB*5
V'?phä5rr
(DH,nabb
ti 'rvcn
9-•cr?H1
rr-l>'O
V+-¤·rb
¢·~Of!]P-I t‘*
(bnr,SS
L6
'89
O'L
2T
90
29
09
99
2°L
S9
‘SC
T2
96
29
28
82
°99
Z'S
8_
22
L2
00
8‘
€T
'8€
Z°O
L0
T9
2S
LL9
9°O
€9
°99
S€€2
OS
LL
2'€
28
°29
29
T2
S2L
99
'LT
T'2
€6
00
2C
CL
98
°2T
L°€
29
98
TS
L9E
9'6
L°L
T8
2L
I0
99
LS
°L6
'€T
CT
9T9
29
66
°SO
°TT
96
9I
00
9‘
LC
'9€
°6_
06
€T
SL9
99
°92
'89
82
TO
SS
OO
'99
'L0
02
T9
29
08
'£O
'L9
TT
I0
09
-9
9°€
L°9
6€
OI
SL9
99
'€€
°9L
96
09
9O
2°€
6°9
20
69
29
90
'€9
'92
98
00
9S
L'?
T°9
06
LS
LZS
S°Z
L'9
29
L0
9€
ST
'ZO
°99
69
92
€L
8'I
_9
°€8
99
C0€
(lua
sla
d)
(O
TX
mg)
(°D3S
)(UO
)u
oyq
ep
pxg
Vu
ya
ga
myl
·dmaL
go
Qu
aq
xg3
q8
;aM
‘
T·60—
S9
'IIa
tdw
vs
JO}
SJIH
SQ
HIP
QU
BLH
IJSÖ
XH
°3A
IEI“I¥IV.L
°9
0'9
96
'TZ
TL
9T
€0
99
66
‘99
9°TC
T8
€6
29
28
99
'99
0°2
9O
TL2
0088
9°9
£9
'L9
{TS
2S
LL6
€'6
22
'99
92
92
09
L2
9'2
2L
°T9
99
T2
92
L9
6°9
T€
'T€
9CO
ZC
CL
99
°2T
8'€
2299T
9L
9L€°C
TT
‘GT
97
lTu
fi.
69
°8C
°9T£09T
92
9»
92
'L9
'ET
989T0
09
N0
£°9
9'T
T29Z
T9
L9
M9
9'9
€°0T992T
0999
L'9
8'9
69TT
92
99
2'9
6'L
TTTT0
09
26
°E2
°L6€O
T9
L9
99
‘€L
°92
96
09
9L
9°E
9'9
20
69
29
29
'€9
'90
98
00
99
9'Z
9'9
99
L9
L€
T9
'€Q
'92
2L
09E6
2°E
T°9
99
99
22
9T
°€—
9'9
99
900E
(QU
BO
JSÄ
)(V
QI
xm
3)(·0
ag
)(3
0)
‘u
oyq
ep
yxou
;23
am
;;_·d
ma
lgo
quaqxgq
qä
yan
2-0
2*7
7T
IBTÖ
MES
19
}S
JIHS
QH
IEJU
SW
TJQÖ
XH
'QA
IH
THV
L
99
*0L
L'lT
T9
99
9C
989
9*9
9L
°26
28
09
92
98
6*6
60
°6L
69
2L
C(L
88
2*9
66
*89
90
92
9L
L9
L*6
€6
*69
90
62
09
L&
0*L
22
°üü
2€
22
92
L9
9*0
29
*29
99
02
OO
L2
L*6
T2
*92
Elö
l9
L9
*0*7
l0
*6T
Tlll
09
99
6°6
L°9
TL
£9
T9
29
A9
9*9
9*E
Tl6
9T
00
9’
26
'LL
°TT
90
6T
9L
9Q
69*99*0T
90
El
0999
0°9
9*6
9lZ
T9
29
69
°90
*69
2T
T0
09
·0
6*9
9*8
96OT
SL6
L2
*9€
°86
L6
09
62
T'9
T*8
6l6
92
62
T*9
l'82
98
00
69
9*6
L*L
66
L9
L€
96
*60
*L9
6L
CS
EE
0*6
6*9
TO
L9
29
99
°€9
*99
99
00
9
‘(1u
a0
1a
d)
(VO
Ix
mä)
(·oa
g)
(30
)u
oyq
ep
yxgu
yeg
am
yl
·dmaL
go
nu
aq
xg
äya
ü
T*T
Z*7
VII
BTÖW
BS103
$3
T“S
9H
lgü
uö
mila
dxä
°TA
IH
TSV
L
TAB
LE1
Vj,
Exp
erim
en
tal
Re
su
lts
for
Sam
ple
IL4
2-0
9-2
We
igh
tE
xte
nt
of
Temp
.Ti
me
Gai
n4
Oxi
datio
n(O
C)
(Se
c.)
(gm
x10
)(P
erc
en
t)
30
06
85
5.8
'3
.68
32
57
28
6.5
4.0
8~
35
07
76
6.9
4.3
53
75
831
7.5
4.7
14
00
88
87
.74
.85
42
595
17
.64
.79
45
010
157
.64
.81
47
51
08
97
.74
.88
~5
00
1167
7.9
4.9
85
25
1253
8.8
5.5
355
0134
59.5
5.99
g57
514
471
0.9
6.8
96
00
1555
12
.57
.90
62
516
731
4.9
9.4
1E
65
01
79
51
8.8
11
.87
675
1935
23
.71
4.9
27
00
20
82
31
.61
9.9
2·
72
52
23
84
0.8
25
.74
75
02
41
35
2.7
33
.24
77
52
59
86
5.2
41
.10
80
02
80
81
79
.34
9.9
98
25
30
43
97
.16
1.1
98
50
3311
11
5.7
72
.92
TAB
LE1
Vk.
Exp
erim
en
tal
Re
su
lts
for
Sam
ple
IL4
5-0
6-2
We
igh
tE
xte
nt
of
Tgmp
.Ti
me
Gai
n4
Oxi
datio
n(
C)
(Se
c.)
(gm
x10
)(P
erc
en
t)
30
06
53
4.7
42
.99
32
56
94
5.4
3.3
93
50
73
95
.63
.53
-3
75
78
76
.03
.76
40
08
40
6.5
4.1
04
25
89
86
.74
.23
14
50
96
26
.94
.37
q4
75
10
30
7.0
4.4
4·
50
01
10
97
.24
.54
°5
25
11
93
7.7
4.8
455
012
628.
75
.4
9~
57
51
38
09
.56
.01
60
01
48
31
0.8
6.8
36
25
15
94
12
.57
.91
65
017
231
5.6
9.8
66
75
18
58
19
.31
2.1
57
00
1995
25
.51
6.0
37
25
21
38
32
.42
0.4
67
50
23
30
41
.42
6.1
47
75
24
86
51
.63
2.5
58
00
26
98
‘6
3.4
39
.99
82
52
92
17
7.2
48
.68
:°j4
·. “8504
3178‘ 93.2
58.771
--
2?
*T9
9*2
II6
2T
20
98
°6
2*2
96
*96
08
82
S28
68°%9
0*2
9·
29
92
00
80
?*9
22
*L9
29
?2
SLL
09
*62
9*9
98
L2
20
9L
29
*22
8°T
?6
IT2
92
L8
6*9
I?
*T2
0L
6I
CO
L9
2*2
TL
*22
?28T9
L9
L2
”6Y
°LI
20
fT099
.2
9'L
L*2
I6
LS
I9
29
V92*9
L*TI69?T
009A
8?
*9I*0
T992T
SL9
qL
9'?
0*6
OLZT
C9S
92
*?I*8
TQTT
92
9II'?
9*L
?0
Il0
09
‘
09
*20
*L2
2O
IS
L?9
8*2
T*L
99
6O
9?2
L*2
6*9
268S
2?L
9'2
9*9
22
8O
0?6
2*2
2*9
08
LS
L26
I*26
*90
2L
09
29
6*2
9*9
?8
99
22
28
*2,
2*9
92
90
02
(qu
aa
xad
)(V
0;
xm
ä)(·a
ag
)(3
0)
uo
pq
ep
pxg
upegam
p;·d
ma
;go
quaqxg1q8paM
I—LQ
—Q
Vj;
ap
dm
eg
10
;sq
pn
sa
gIe
qu
am
pxa
dxg
'IA1
HIQ
VL
TAB
LEIV
m.
Exp
erim
en
tal
Re
su
lts
for
Sam
ple
IL4
4-1
9-1
We
igh
tE
xte
nt
of
Tgmp
.Ti
me
Gai
n4
Oxi
datio
n(
C)
(Se
c.)
(gm
x10
)(P
erc
en
t)
30
06
43
6.4
”3
.49
32
56
86
6.9
3.7
23
50
72
57
.54
.06
37
57
78
7.6
4.1
04
00
83
07
.64
.12
42
58
88
7.4
4.0
14
50
95
0_
7.6
4.1
44
75
1015
7.8
4.2
44
-5
00
10
92
8.0
4.3
35
25
11
76
9.0
4.8
65
59
1260
10
.15
.47
Q5
75
13
54
11
.56
.25
_6
00
1452
13
.17
.12
625
15
58
15
.58
.42
n
65
01
68
21
8.4
9.9
96
75
1812
21
.61
1.6
97
00
19
49
26
.91
4.5
77
25
20
95
35
.81
9.4
1-
75
02
25
84
8.9
26
.51
—7
75
24
34
61
.83
3.4
38
00
26
14
‘7
4.0
40
.09
82
52
83
99
0.0
48
.74
85
03
05
81
06
.55
7.6
8
LT
'99
T'2
0T
22
22
09
8T
9'9
9?
'68
L9
62
92
86
?°L
?T
'9L
·L
2L
20
08
02
'62
0'Z
9T
?929
LL
?9
'T2
O'0
96
92
2G
SL
6?
'?2
L'@
2L
6T
29
?L
68
‘LT
2'8
29
90
2C
OL
L9
'2T
9'T
22T
6T9L9
L9
'0I
$°9
T9G
LT0
99
L2
'8Z
'2T
299T9
29
_T
6'9
6'O
T8?9T
00
92
0'9
9'6
L2?TS
L90
9'9
L'8
L22T0
99
T6
'?8
'L9?2T
92
9S
9'?
?'L
29TT
00
9T
T9
'?T
'LT
80TE
L?S
9'?
0'L
2TOT
09
?L
T'?
9'9
2?
6Q
2?L
T'?
9'9
288O
0?9
6'Z
2°9
92
8Q
L22
L'2
6'9
2LLC
S2
‘6
€'€
?'9
92
L9
22
L2
'€2
'92
89
00
2<¤¤@¤¤=->a>
<ot
Xul?)
<·¤¤s>(SO)
uo
yjg
ep
jgxg
,7u
‘gE
QB
lllll°d
u1
3]„
JOQ
UBQXH
JLIQFG
M
aydmeg
10
;sq
ynsa
g{¤
3u
am
;1a
dxg
·uA
Igqgvg
TAB
LEIV
O.
Exp
erim
en
tal
Re
su
lts
for
Sam
ple
IL4
2-0
8-2
We
igh
tE
xte
nt
of
Tgm
p.Ti
me
Gai
n4
Oxi
datio
n(
C)
(Se
c.)
(gm
x10
)(P
erc
en
t)
30
06
78
4.6
”2
.20
32
57
20
5.1
2.4
13
50
767
5.3
2.5
23
75
81
86
.12
.88
40
087
16
.73
.18
42
59
30
6.7
3.1
84
50
99
67
.03
.34
475
1067
7.1
3.3
9·
·5
00
1143
7.8
3.7
0-
52
51
22
68
.43
.97
550
1313
9.24.3
7g
57
51
41
61
0.4
4.9
5p
60
01
51
51
2.2
5.8
16
25
1637
14
.56
.90
i
65
017
521
8.8
8.9
46
75
18
84
24
.61
1.6
67
00
20
22
34
.31
6.2
8·
—7
25
21
75
47
.52
2.5
67
50
23
40
60
.82
8.8
877
52
52
77
5.9
36
.03
80
02
72
9‘
92
.04
3.6
78
25
29
65
11
2.2
53
.24
85
03
21
91
32
.36
2.7
9
TAB
LEIV
p.
Sum
mar
yo
fE
xp
erim
en
tal
Re
su
lts
Initia
lS
ampl
eO
xid
atio
nS
ampl
e_A
Wei
ght
(mg)
Pa
rtic
leS
ize
at
850O
C(Z
)
IL4
5-1
0-2
35
0.1
.4
5-7
5um
79
.70
IL4
4-2
2-2
30
0.0
.4
5-7
5um
80
.72
IL4
4-2
2-1
35
0.0
45
-75
um7
6.4
5IL
42
-10
-13
00
.24
5-7
5um
83
.70
IL4
2-1
1-1
39
9.0
45
-75
um8
7.0
3
IL4
5-0
8-1
30
0.6
75
-15
0um
65
.82
IL4
5-0
9-1
34
9.3
75
-15
0um
68
.97
IL4
4-2
0-2
35
0.0
75
-15
0um
66
.05
IL4
4-2
1-1
29
9.8
75
-15
0um
70
.65
‘IL
42
-09
-23
00
.97
5-1
50
um7
2.9
2
IL4
5-0
6-1
30
0.8
15
0-1
80
um5
8.7
745
IL4
5-0
7-1
35
0.7
15
0-1
80
um·
61
.43
U1IL
44
-19
-13
50
.11
50
-18
0um
57
.68
IL4
2-0
8-1
30
0.0
15
0-1
80
um6
5.1
7IL
42
-08
-23
99
.61
50
-18
0um
62
.79
1CU.
CD
+
xz
LEG
EN
DO
8000
j{jeA
1-a 1-
ZEm
1¤.=1
5-10-2
CI
+Q
Q*
AIL
44-
22-2
,.‘
BD-Ü
ÜX
élA
IL4
4-22
-1O
jf·
IL4
2-10
-1|—
„+<>
¤+
IL4
2-1
1-1
gi
Öm
¤J_
A
45-75
zuOU
M 1tra
um
{IJA
U]72
[1-_C
D14
40.C
D2.L
BO.G
]2
H.
CDBH
II.00
T1
NE
(SE
C)
.F
igkl
fé6-
The
pe
rce
nta
ge
of
the
ore
tica
lo
xid
atio
nve
rsu
stim
efo
rth
e4
5-7
5um
pa
rtic
les.
1CIl0
0
00LE
GEN
ÜI-
]+
Emg
*m
1L6
5-08
-1Q
Sam
+A
1L‘+
5-09
-1x
QÜé
A1L
66-
20-2
OA
„IL
44-2
1-1tm
001
+IL
‘+2-
O9-2
5,M
1*E
#9äl
zuUO
ip.}75
-150
,d>M
1CRO
Naß
PF
IRT
I[L
ES
.CU
.'-et
wm
vInI
VCU
720.
ED1+
*+0.0
021
80.0
02H
!.00
3800
. 00
T1
ME
(5E
CJ
·F
igu
re7.
The
pe
rce
nta
ge
of
the
ore
tica
lo
xid
atio
mve
rsu
stim
efo
r7
5-1
50
ump
art
icle
s.
lCU.(D
u
äm
mLE
GEN
D•-4
.I-
·IL‘I*5-06
2CK
nQ
A*
•-lL
‘IS
•D7
-lH
Sum*°m
• IL‘I
~—
Oto
m•
IL4
2-08
-1I"
:‘”·
IL‘I
2-D
8-2
§>Z
4111]]
AuÜLI
JO-
L.!E+
2'°°
150.1
60mw
r~111:60~_
PFIR
TIELE
S.¤.'J
(IJ72IJ.C
U1.440
.11'ZLEC
LCIJ2£l.
CD3&'D
.CIJTIM
E(SEE
)F
igu
re8
.T
hep
erc
en
tag
eo
fth
eo
retica
lo
xid
atio
nve
rsu
stim
eo
f1
50
-18
0um
pa
rtic
les.
49
before the sample reached 30000, but the stable value varied by
i_0.25 mg. The four standards were averaged together after it was
shown that their behavior did not correlate with their weight. ,
The charts obtained from the oxidation of ilmenite were actually
records of two simultaneous processes——the steady oxidation process
superimposed on the somewhat erratic behavior of the standard. A
j_O.25 mg discrepancy is the same as adding or subtracting as much
as i 3 percent to the obtained results. This quantity will be taken
as the upper limit of experimental error.
One outstanding feature of the graphs of oxidation is their
failure to reach completion. A series of l gram samples were held
at temperatures from 500 to 80000 for twelve hours to determine if
the reaction will reach its theoretical limit of 5.27 percent weight
increase. Table Va shows that at low temperatures the oxidation was
very sluggish, with the process showing a pronounced particle size
effect. At the higher temperatures, the specimens oxidized very
rapidly, but still were not completely reacted at the end of twelve
hours. Samples weighing 300 mg. were oxidized at 75000 for 24 hours
(see Table Vb) with no improvement shown over the results at twelve
hours. It was noted that some sintering had taken place during the
longer times, especially in the very fine particles. This could
explain why the smallest particles did not oxidize more thoroughly.
A major source of concern was the inability to obtain reproducible
heating rates with the equipment used (Figure 9). The heating rate
varied due to changes in ambient temperature. Careful consideration
50
TABLE Va. Weight Gained After 12 Hours at Temperature
150-180 um Particles 75-150 um Particles
Weight Fraction Weight FractionTempgrature Increase Reacted Increase Reacted
( C) (Percent) (Percent) (Percent) (Percent)
500 2.24 42.4 2.16 41.0
600 3.08 58.4 4.39 82.3
700 4.70 89.2 4.80 91.0
800 4.94 93.7 5.00 94.8
TABLE Vb. Weight Gained After 24 Hours at 750OC
Weight FractionIncrease Reacted
Particle Size (Percemt) _ (Percent)
45-75 um 4.70 89.1
75-150 um 4.76 90.4
150-180 um 4.55 86.4
l.(]]].Ü
J
EE
lm
au
.ml
A-?
A1
—·
1
¤F
X•
2Lu
+m.m
•-—
I DZ
ILIIJ
F
.CD
CD7Z
J.CD
.L‘H
G.CD
BEIL
CIJ
TIM
E(S
EE
)F
igu
re9
.A
plo
to
fte
mp
era
ture
ve
rsu
stim
efo
rth
ekin
etic
sam
ple
s.
52
of the reaction process is necessary to justify the use of thermo-
gravimetric techniques employing non—linear heating rates.
. The rate of oxidation is controlled by either activation energy
considerations and by the rate of supply of the reactants. These
cases may be considered:
(l) if the reactants are finely interspersed, say on an atomic scale, ‘
the reaction will be controlled by the energy available. If it
is less than the activation energy of the process, no reaction
will occur. If it is greater, the process will occur
instantaneously.
(2) If the reactants are not interspersed, as in a solid—gas reaction,
and energy is available in excess of the activation energy of the
process, then the rate of reaction is controlled only by the
supply of any one of the reactants. Consider:
(a) if the product forms a coherent layer on the surface of a
non—porous solid, the diffusion of one of the reactants
through this layer will be the rate controlling step, and
the rate will be proportional to the inverse square of the
product thickness.
(b) if the product layer is not adherent to the surface, and
diffusion is slow with respect to the reaction rate, the
interfacial area of the shrinking particle becomes rate
controlling.
For isothermal experiments, a log-log plot of l—(l—X)l/3versus
time can be made, where X is the fractional amount of reaction that
53
has occurred. The slope of the plot determines the reaction mechanism.
No such convenient method exists for the determination of kinetic
data when the temperature is varying as the reaction occurs(33).
The method of Chatterjee(29) was first attempted, but his method
was apparently not well suited to this experiment, nor was the method
employed by Freeman and Carrol1(28).i The reasons for this unsuit-
ability have not been discovered. It is speculated that the major
difference is in the type of sample used. When deriving their methods,
both investigators employed cylindrical pellets rather than loose
powder. lt is not clear why this should have an effect, but results
using these methods, even on "smoothed" data, were very erratic.
Chatterjee's method is particularly suspect, as it requires two
different sample weights. When powdered ilmenite was used, no sample
l height effect was demonstrated. Shah and Kha1afalla(3l), who employed
Chaterjee‘s method, had cylindrical pellets whose heightzdiameter
ratio varied from 1:1 to 1:3 as the weight was varied. The effect of
this difference in geometry cannot be explained by their report.
Other methods(26’27) required that a linear temperature increase be
employed so that a transformation of variables could take place.
These techniques obviously were not applicable.
The following method of data analysis was employed:
(1) using piecewise—linear and piecewise—polynomial curve—fitting
techniques, a least squares curve was determined for each
' of the three size fractions (Figure 10).
Attention Patron:
Page Üji omitted from
numbering
55
(2) the slopes of the curves were determined by numerical
differentiation of the "best-fit" curve. The slope was
equivalent to the reaction rate, expressed as the rate of
change in the percentage of the particle that had undergone
oxidation versus time.
(3) two assumptions were made at this point: (a) the particles
were spherical with a diameter equal to the mean of the
size distribution, and (b) the product layer formed uniformly
on the surface of the sphere. Adherence of the product
layer is not implied.
(4) the radius of the unreacted core of the spherical particle ·
was calculated by:
ri = r<1where 6i is the fractional oxidation.
(5) the thickness of the product layer, if adherent and un-
fractured, was calculated by
ti = (r0—ri)(l.07)
This was based on a seven percent lattice expansion upon the
completion of the reaction
4FeTi03 + 02 + 2Fe203 + 4TiO2
(6) The interfacial area was based on the spherical model, and
calculated by
ai = 4wri2
56
(7) the reaction mechanism would be determined from Arhenius
plots of the normalized rates of reaction versus temperature.
This assumes a governing relation of the form
Ri = RO€—Q/RTi,
where: Ri is the normalized rate at temperature Ti (OK),
RO is the initial normalized rate, Q is the activation energy,
R is the ideal gas constant, and T is absolute temperature.
(8) The normalization would involve nothing in the case of
chemical reaction control, division by the instantaneous
interfacial area if the adsorption of fluid reactants at the
reaction interface is controlling, or multiplication by the
square of the product thickness if diffusion is the rate
determining step.
(9) Ideally, the data, if properly normalized, should fall on
the same linear curve.
The data used in the making of Figures 11-13 are presented in
Tables VIa—c. Inspection shows that the best agreement is obtained
from Figure 13, the plot in which the reaction rate was normalized
to the assumed product layer thickness. The agreement is particularly
good at temperatures above 525OC. The activation energies for the
two sections of the curve are determined from the slope of the
Arrhenius plots, as below:
R = R e‘Q/RTo
Taking the logarithm of both sides, we obtain:V
-1.tI
J
°_°am
mmcz
E_L
4Ügg
gggg
LEG
EN
DÄ
tm
Züm
‘45
-75
NIC
.Q
L80
¤¤•
75-15
0m
t.O
°bm
•15
0-18
0N
IC.
r-Gm
ZS
AI
Ln
H-2
209A
“
Egm
mm
mm
mm
mca
QA
AA
AA
AA
AA
E_2
B06
66
66
66
66
“‘e LD C) —'
-3.::
1.7
51.
CI]
1..25
1..E
l1.
752
(Il
1.D
DÜ
/T(°
K)
Fig
ure
ll.
Ap
lot
of
oxid
atio
ura
te(lo
glO
d(%
ox)/
dt)
vers
us
recip
roca
la
bso
lute
tem
pe
ratu
re.
3.Ü3
‘EI E1
220
@0%
LEG
EN
DE1
"‘
El.
CKAA
mm¤
45-7
5M
IC.
Agg
140
°A¤A
wmA
75-15
0M1
:-,
QCf
AÜ
A1
50
-18
0N
IC.
LL;
°AA
EIE
DE
IIIIE
IEIE
IEIE
IEI
I-_
°©A
U,G
:A
Aoo
Q.9
3°’©
A-J
OQQ
AA
AA
AA
AA
Ag-
Q—*
-.20
-L.¤
3 „.7
51.
CD*
1.25
1.50
1.75
2.00
10
00
/T(°
K)
Fig
ure
12.
Ap
lot
of
oxid
atio
nra
te(n
orm
aliz
ed
toin
terf
acia
la
rea
)ve
rsu
sre
cip
roca
la
bso
lute
tem
pe
ratu
re.
-5.C
D
¢-6.
ZUAQ,
L EG
END
uno
Nm
p0
45-7
5M
IC.
*—‘°^
°•
75-1
60M1
1:.-7
.10°°
•X
Btw
•1S
Ü-18
0N
IC.
LL]
EJ»—
3*..
Q:AQ
xo
(Iuf
¤^X
30
X-9
.8]
mm
XX
gg
XX
°°°
ÜB
ÜÜ
Ü
-L1.
c¤.7
51.
CI]
LZ
S1.
5D1
.75
ZG
]
LÜÜ
Ü/T
(°K
)F
igu
re13
.A
plo
to
fo
xid
atio
nra
te(n
orm
aliz
ed
top
rod
uct
laye
rth
ickn
ess
sq
ua
red
)V
EY
SU
SI°
€C
Ä1
p][°O
Ca
l8
bS
OlU
t€.
t€lT
1p
€l.°3
tUI°
ۥ
TAB
LEV
la.
Ca
lcu
late
dR
esu
lts
for
the
Oxid
atio
no
f4
5-7
5pm
Pa
rtic
les
(de
term
ine
dfr
om
lea
st
sq
ua
res).
Exte
nt
of
Inte
rfa
cia
lP
rod
uct
Tem
p.T
ime
Oxid
atio
nO
xid
atio
nR
ate
Are
aT
hic
kn
ess
(OC
)(S
ec.)
(Pe
rce
nt)
(Pe
rce
nt/S
ec
.)(s
q.c
m.)
(cm
.)
300
66
03
.29
4.5
88
5-0
3_
4.4
24
5-0
46
.98
55
-05
32
57
04
3.4
94
.58
85
-03
4.4
18
5-0
47
.42
15
-05
35
07
49
3.7
04
.58
85
-03
4.4
12
5-0
47
.86
55
-05
37
58
01
3.9
44
.58
85
-03
4.4
04
5-0
48
.37
85
-05
40
08
54
4.1
84
.58
85
-03
4.3
97
5-0
48
.91
05
-05
42
59
14
4.4
54
.58
85
-03
4.3
89
5-0
49
.49
75
-05
45
09
79
4.7
54
.58
85
-03
4.3
79
5-0
41
.01
55
-04
47
51
05
05
.08
4.5
86
5-0
34
.36
95
-04
1.0
86
5-O
45
00
11
27
5.4
34
.58
85
-03
4.3
58
5-0
41
.16
35
-O4
52
51
20
85
.89
4.4
83
5-0
34
.34
55
-04
1.2
61
5-0
455
01
29
96
.46
6.5
73
5-0
34
.32
75
-04
1.3
88
5-0
457
51
39
87
.26
9.5
76
5-0
34
.30
25
-04
1.5
63
5-0
46
00
15
04
8.4
71
.35
05
-02
4.2
65
5-O
41
.83
25
-04
625
16
17
10
.27
1.8
49
5-0
24
.20
85
-04
_2
.23
65
-04
65
01
74
01
2.9
32
.48
85
-02
4.1
25
5-0
42
.84
25
-04
67
51
07
41
6.7
93
.29
75
-02
4.0
02
5-0
43
.74
45
-04
700
20
16
22
.17
4.2
84
5-0
23
.82
35
-04
5.0
49
5-0
472
52
16
72
9.3
64
.99
65
-02
3.5
88
5-0
46
.89
25
-04
750
23
36
37
.66
4.9
96
5-O
23
.30
15
-O4
9.1
81
5-0
477
52
52
04
6.8
64
.99
65
-02
2.9
68
5-0
41
.19
75
-O3
80
027
215
6.9
14
.99
65
-O2
2.5
81
5-0
41
.54
15
-03
825
29
48
68
.27
4.9
96
5-0
22
.10
55
-04
2.0
03
5-0
3-
850
32
04
81
.02
4.9
96
5-0
21
.49
45
-04
2.6
79
5-0
3
TAB
LEV
lb.
Ca
lcu
late
dR
esu
lts
for
the
Oxid
atio
no
f7
5-1
50
umP
art
icle
s(d
ete
rmin
ed
fro
mle
ast
sq
ua
res).
Ext
en
to
fA
Inte
rfa
cia
lP
rod
uct
Tem
p.T
ime
Oxid
atio
nO
xid
atio
nR
ate
Are
aT
hic
kn
ess
(OC
)(S
ec.)
(Pe
rce
nt)
Pe
rce
nt/S
ec.)
(sq
.cm
.)(c
m.)
30
06
60
3.3
63
.30
55
-03
’1.5
55
5-0
31
.33
75
-04
A»
32
57
04
3.5
03
.30
55
-03
1.5
53
5-0
31
.39
65
-04
35
07
49
·3.6
53
.30
55
-03
1.5
51
5-O
31
.45
65
-O4
37
58
01
3.8
23
.30
55
-03
1.5
50
5-0
31
.52
55
-04
40
08
54
4.0
03
.30
55
-03
1.5
48
5-0
31
.59
75
-04
42
59
14
4.2
03
.30
55
-03
1.5
46
5-0
31
.67
65
-04
45
09
79
4.4
·13
.30
55
-03
1.5
43
5-0
31
.76
45
-04
47
51
05
04
.65
3.3
05
5-0
31
.54
15
-03
1.8
59
5-O
45
00
1127
4.9
03
.30
55
-03
1.5
38
5-0
31
.96
35
-04
525
12
08
5.1
65
.35
15
-03
1.5
35
5-0
32
.06
75
-04
550
12
99
5.8
27
.37
45
-03
1.5
28
5-0
32
.33
65
-04
E157
51
39
86
.58
8.2
04
5-0
31
.52
05
-03
2.6
50
5-0
46
00
1504
7.5
41
.00
85
-02
1.5
09
5-0
33
.04
75
-04
625
16
17
8.8
41
.31
95
-02
1.4
95
5-0
3l
3.5
90
5-0
465
01
74
01
0.7
41
.78
95
-02
1.4
74
5-0
34
.39
05
-04
67
51
87
41
3.5
72
.45
75
-02
1.4
43
5-O
35
.60
35
-04
700
20
16
17
.67
3.3
44
5-0
21
.39
75
-03
7.4
13
5-0
472
52
16
72
3.5
04
.18
85
-02
1.3
30
5-0
31
.00
95
-03
75
02
33
63
0.5
34
.18
85
-02
1.2
48
5-0
31
.35
15
-03
775
25
20
38
.24
4.1
88
5-0
21
.15
35
-03
1.7
53
5-0
380
02
72
14
6.6
64
.18
85
-02
1.0
46
5-0
32
.23
35
-O3
825
29
48
56
.19
4.1
88
5-0
29
.17
55
-04
2.8
41
5-0
385
;)3
20
46
6.8
74
.18
85
-02
7.6
14
5-0
43
.63
95
-03
TA
BLE
Vlc
.C
alc
ula
ted
Re
su
lts
for
the
Oxid
atio
no
f1
50
-18
0um
Pa
rtic
les
(de
term
ine
dfr
om
‘
lea
st
sq
ua
res).
Ext
en
to
fIn
terf
acia
lP
rodu
ctT
emp.
Tim
eO
xid
atio
nO
xid
atio
nR
ate
Are
aT
hic
kne
ss(O
C)
(Se
c.)
(Pe
rce
nt)
(Pe
rce
nt/S
ec.)
'(s
q.c
m.)
(cm
.)
300
66
03
.16
2.5
59
5-0
33
.34
95
-03
1.3
47
5-0
432
57
04
,3
.28
2.5
59
5-0
33
.34
65
-03
1.9
14
5-0
435
07
49
3.3
92
.55
95
-03
3.3
43
5-0
31
.98
25
-04
37
580
13
.53
2.5
59
5-0
33
.34
05
-03
2.0
61
5-0
44
00
85
43
.66
2.5
59
5-0
33
.33
75
-03
2.1
42
5-0
442
59
14
3.6
12
.55
95
-03
3.3
34
5-0
32
.23
25
-04
45
09
79
3.9
82
.55
95
-03
3.3
30
5-0
32
.33
15
-04
475
10
50
4.1
62
.55
95
-O3
3.3
26
5-0
32
.43
95
-04
50
0·
1127
4.3
62
.55
95
-03
3.3
21
5-0
32
.55
65
-04
.·
525
12
08
4.6
24
.08
75
-03
3.3
15
5-0
32
.71
25
-04
550
1299
5.2
05
.93
45
-03
3.3
02
5-0
33
.05
65
-04
Q57
513
985
.84
7.1
11
5-0
33
.28
75
-03
3.4
39
5-0
4-
60
01
50
46
.70
9.3
28
5-0
33
.26
75
-03
43
.95
95
-04
625
1617
7.9
41
.27
65
-02
3.2
38
5-0
34
.71
25
-04
650
17
40
9.8
11
.78
25
-02
3.1
94
5-0
35
.85
95
-04
67
51
87
41
2.6
42
.48
25
-02
3.1
26
5-0
37
.63
35
-04
700
20
16
16
.80
3.3
98
5-0
23
.02
65
-03
1.0
30
5-0
372
52
16
72
2.8
53
.86
15
-02
2.8
78
5-0
31
.43
55
-03
.7
50
23
36
29
.45
3.8
61
5-0
22
.71
15
-03
1.9
02
5-0
377
52
52
03
6.5
63
.86
15
-02
2.5
26
5-0
32
.43
95
-03
800
27
21
44
.33
3.8
61
5-0
22
.31
55
-03
3.0
73
5-0
382
52
94
85
3.1
13
.86
15
-02
2.0
65
5-0
33
.86
65
-03
-85
03
20
46
2.9
73
.86
15
-02
1.7
64
5-0
34
.88
35
-03
63
ln R = ln RO — (%) é
This shows that the slope of the Arrhenius plots should be equal to
the activation energy divided by the ideal gas constant. The slopes
were obtained by linear regression using the calculated data that
are given in Table VII. The activation energy of the process in the
initial stage (300-525OC) of reaction is 4.35 i_0.5 kcal/mole, and
is 50.2 i 3.0 kcal/mole at higher temperatures.
In the initial stage of reaction, many features on the surface
oxidize more easily than the sample as a whole. Evidence for this
is two—fold: the low activation energy for the process, and the
lack of a good fit to a single straight line. Diffusion control
according to the assumed model does not begin until the interface
begins moving uniformly toward the particle center.- The activation
energy obtained for the initial stage of oxidation is not generally
applicable to other kinetic techniques.
Above 60006, a very good fit to the assumed diffusion model
is obtained. This fit could have been improved by a more quantitative
determination of the mean particle diameters. The activation energy
is of a magnitude similar to that of the diffusion process in many
metals (40-60 kcal). This may be taken as confirmation of the
suggestion(l8) that iron is diffusing through the particle to the
reaction zone.
64
TABLE VII. Calculated Results Used in Determination of ActivationEnergies by the Method of Linear Regression.
nn Tux (Rate x Thicknessz)
Activation45-75 um 75-150 um 150-180 um Energy
T(OC)1000/T(OK) Particles Particles Particles (kcal/mole)
300 1.745 24.523 23.552 23.1621
325 1.672 24.402 23.466 23.090
350 1.605 24.285 23.382 23.021
375 1.543 24.159 23.289 22.942
400 1.486 24.036 23.197 22.865 0*35 i 0*5425 1.433 23.908 23.100 22.783
450 1.383 23.775 22.998 22.696
475 1.337 23.640 22.893 22.606
500 1.294 23.503 22.784 22.512*J525 1.253 23.364 22.199 21.965
550 1.215 22.790 21.608 21.314
575 1.179 22.176 21.275 20.896—7
600 1.145 21.515 20.790 20.343
625 1.114 20.802 20.193 19.680
650 1.083 20.025 19.486 18.912 ~
675 1.005 19.193 18.680 18.052
700 _ 1.028 18.333 17.812 17.1385O°2 i‘3°O
725 1.002 17.556 16.971 16.347
750 0.978 16.983 16.387 15.784
775 0.954 16.452 15.866 15.287
800 0.932 15.947 15.382 14.824
825 0.911 15.423 14.900 14.365
850 0.890 14.841 14.405 13.898J
CONCLUSIONS
(l) The oxidation of ilmenite has been shown to obey the followingreactions:(a) below 800OC, ilmenite oxidizes to form pseudorutile and
hematite by the reaction ·
l2FeTi03 + 302 + 4Fe2Ti304 + 2Fe203
(b) above 800OC, pseudobrookite and rutile are formed according
to the reaction
4FeTi03 + 02 + 2Fe2Ti05 + 2Ti02
The presence of pseudorutile in a specimen heated twelve hours
at 850OC indicates that it may be an intermediate compound in
the reaction.
(2) The change of oxidation mechanism above 80OOC should not affect
an industrial process because no definite crystal structure is
apparent at the end of the short periods of time needed to react
finely powdered ilmenite to its apparent limit of oxidation. Also,
the theoretically required amount of oxygen is unaffected by the
change in reaction mechanism.
(3) The oxidation of ilmenite is diffusion controlled, with an
apparent activation energy of 50 kcal/mole above 550OC. The
diffusion of iron is considered to be the controlling step in
this process.
(4) Thermogravimetric analysis does not require a particular variety
of heating rate to give useful results. The ability to accurately
model the experimental process is still the limiting factor.
I
65
66
(5) The rate of oxidation at 700OC is more than an order of magnitude
higher than the rate at SOOOC. This is an important factor in
the design of any industrial process.
RECOMENDATIONS FOR FURTHER STUDY
Of importance to any industrial process using the oxidation of
ilmenite as a preliminary step is the ease with which the iron and
titanium can be separated. The ferrous iron present in ilmenite is
converted to ferric iron by either of the reactions shown to occur,
but the oxidized ilmenite does not immediately assume any identifiable
structure.
Research into the extraction of iron or titanium oxides from
this 'amorphous' oxidation product should be carried out. Chemicall
leaching and solid state reduction are two candidates for the process.
67
BIBLIOGRAPHYV
l. Kothari, N. C., "Recent Developments in Processing Ilmenite forTitanium", International Journal for Mineral Processing,_2,pp. 287-305 (1975).
2. Hoch, M., "Critical Review: Winning and Refining", Proceedingsof the Second International Conference on Titanium Science andTechnology, Vol. 1; edited by R. I. Jaffee and H. M. Burte,Plenum Press, New York, pp. 205-231 (1972).
3. Poggi, D., Charrette, G. G., and Rigaud, M., "Reducation ofIlmenite and Ilmenite Ores", Proceedings of the Second Inter-national Conference on Titanium Science and Technology, Vol. 1;edited by R. I. Jaffee and H. M. Burte; Plenum Press, New York,pp. 205-231 (1972).
4. Fetisov, V. B., Leont'ev, L. I., Kudinov, B. Z., and Ivanova, S. V.,Characteristics of the Reduction of Oxidized Ilmenite Concentrates",Russian Metallurgy (Metally) 2, pp. 18-23) (1968).
5. Khairy, E. M., Hussein, M. K., and E1-Tawil, S. Z., "Preoxidationof Ilmenite Ores and its Bearing on Their Solid State Reductionwith Hydrogen", National Metallurgical Laboratory TechnicalJournal (Jamshedpur, India) 8, pp. 10-14 (1966).
· 6. Sinha, H. N., "Murso Process for Producing Rutile Substitutes",Proceedings of the Second International Conference on TitaniumScience and Technology, Vol. 1; edited by R. I. Jaffee andH. M. Burte; Plenum Press, New York, pp. 234-245 (1972).
7. Levitskii, V. A., Popov, S. G., and Ratiani, D. D., "Thermo-dynamic Properties of Binary Oxide Systems at ElevatedTemperatures. I1. Thermodynamic Functions of Iron (I1) Titanate(FeTiO ) from Equilibrium and Electromotive Force Data", RussianJournal of Physical Chemistry,_46, pp. 294-296 (1971).
8. MacChesney, J. B., and Muan, A., "Studies in the System Iron Oxide-Titanium Oxide", American Mineralogist, 44, pp. 926-945 (1959).
9. Taylor, R. W., and Schmalzreid, H., "The Free Energy of Formationof Some Titanates, Silicates and Magnesium Aluminate from Measure-ments made with Galvanic Cells Involving Solid Electrolytes",Jorunal of Physical Chemistry, 68, pp. 2444-2449 (1964).
10. Taylor, R. W., "Phase Equilibria in the System Fe0-Fe 0 -Ti02 at13000C", American Mineralogist, 48, pp. 1016-1030 (19647.
68
69
11. Lindsley, D. H., "Iron—Titanium Oxides", Carnegie Institution ofWashington Yearbook, 1964, pp. 144-148 (1965).
12. Akimoto, S., Nagata, T., and Katsura, T., "The TiFe 05—Ti2FeO5Solid solution Series", Nature, 179, pp. 37-38 (1957).
13. Webster, A. H.Ö and Bright, N. F. H., "The System Iron-Titanium-Oxygen at 1220 C and Oxygen Partial Pressures between 1 Atm.and 2 x 10- Atm.", Journal of the American Ceramic Society,44, p. 110-116 (1961). .
14. Lindsley, D. H., "Investigations in the System Fe0-Fe20 —TiO ",Carnegie Institution of Washington Yearbook, 1961, pp. 100-106(1962).
15. Grieve, J., and White, J., "The System FeO—TiO ", Journal of theRoyal Technical College (Glasgow), 4, pp. 441-448 (1939).
16. MacChesney, J. B., and Maun, A., "Phase Equilibria at LiquidusTemperatures in the System Iron Oxide-Titanium Oxide at LowOxygen Pressures", American Mineralogist, 45, pp. 572-582 (1961).
17. Rao, D. B., and Rigaud, M., "Oxidation of Ilmenite and the ProductMorphology", High Temperature Science,_5, p. 323-341 (1975).
18. Temple, A. K., "Alteration of Ilmenite", Economic Geology, 51,pp. 695-714 (1966).
419. Rao, D. B., and Rigaud, M., "Kinetics of the Oxidation of Ilmenite",
Oxidation of Metals, 5, pp. 99-116 (1975).
20. Grey, I. E., and Reid, A. F., "Shear Structure Compounds (Cr,Fe)2Ti _ 0 _ Derived from the a-PbO2 Structural Type", Journal ofSoIiä Sgafe Chemistry, 4, pp. 186-194 (1972).
21. Ramdohr, P., "Important New Observations on Magnetite, Hematite,Ilmenite, and Rutile", Chemical Abstracts, 55, col. 3564 (1941).
22. Curnow, C. E., and Parry, L. G., "0xidation Changes in NaturalIlmenite", Nature, 172, p. 1101 (1954).
23. Teufer, G., and Temple, A. K., "Pseudorutile — A New MineralIntermediate Between Ilmenite and Rutile in the Alteration ofIlmenite", Nature, 211, pp. 179-181 (1966).
24. Overholt, J. L., Vaux, G., and Rodda, J. L., "The Nature ofArizonite", American Mineralogist, 55, pp. 117-119 (1950).
70
25. Karkhanalwala, M. D., and Momin, A. C., "The Alteration ofIlmenite", Economic Geology, 66, pp. 1095-1102 (1959).
26. Kubaschewski, 0., and Hopkins, B. E., Oxidation of Metals andAlloys, 2nd Edition, Butterworths, London, pp. 199-201 (1962).
A27. Wood, G. C., Techniques of Metals Research, Vol. 4, Physico-
chemical Methods in Metals Research, edited by R. A. Rapp,Part 2, Chapter 10A, "Oxidation and Corrosion", Interscience,New York, pp. 509-511 (1970). ·
28. Freeman, E. S., and Carroll, B., "The Application of Thermo-analytical Techniques to Reaction Kinetics. The ThermogravimetricEvaluation of the Kinetics of the Decomposition of CalciumOxylate Monohydrate", Journal of Physical Chemistry, 60, pp.394-397 (1958).
29. Chatterjee, P. K., "Application of Thermoanalytical Techniquesto Reaction Kinetics", Journal of Polymer Science, éä, pp.4253-4262 (1965).
30. Baur, J. P., Bridges, D. W., and Fassell, W. M., Jr., "HighPressure Oxidation of Metals: Oxidation of Metals Under Conditionsof a Linear Temperature Increase", Journal of the ElectrochemicalSociety, 102, pp. 490-496 (1955).
31. Shah, I. D., and Khalafalla, S. E., "Kinetics of Thermal Decom-position of Copper (II) Sulfate and Copper (II) Oxysulfate",U.S. Bureau of Mines Report of Investigation No. 7638 (1972).
32. Themelis, N. J., and Gauvin, W. H., "A Generalized Rate Equationfor the Reduction of Iron", Transactions of the MetallurgicalSociety of the AIME, 227, pp. 290-300 (1963).
33. Wen, C. Y., "Noncatalytic Heterogeneous Solid-Fluid ReactionModels", Industrial and Engineering Chemistry, 60, pp. 34-54(1968).
34. Darken, L. S., and Gurry, R. W., "The System Iron—Oxygen. I. TheWüstite Field and Related Equilibria", Journal of the AmericanChemical Society, 6l, pp. 170-173 (1963).
35. Schwerdtfeger, K., and Turkdogan, E. T., "Techniques of MetalsResearch", Vol. 4, Physicochemical Methods in Metals Research,edited by R. A. Rapp, Part 1, Chapter 4, Chemical Equilibria. B.Equilibria and Transport Phenomena Involving Gas Mixtures andCondensed Phases, Interscience, New York, pp. 322-330 (1970).
36. Bailey, S. W. Cameron, E. N., Spedden, H. R., and Weege, R. J."TheAlteration of Ilmenite in Beach Sands", Economic Geology, 01,pp. 263-279 (1956).
THE OXIDATION OF ILMENITE: A KINETIC STUDY
by
Lawrence J. Corsa, III
(ABSTRACT)I
Synthetically obtained ilmenite (FeO·Ti02) particles were oxidized
under conditions of a non-linear temperature increase over the range
3000-850OC. Short—circuit diffusion in the initial stages of
oxidation gave an activation energy of 4.35 j_0.4 kcal/mole, while
above 525OC the activation energy was 50.2 i_3.0 kcal/mole. The process
was shown to be diffusion controlled, with iron the likely mobile
species above 52500, while no mechanism could be singled out for the
controlling step in the early stages.i
p The oxidation products in O2 after 24 hours were shown to be
pseudorutile (Fe2Ti309) and hematite (Fe2O3) in a finely dispersed
phase at temperatures below about 8000C, with pseudobrookite (Fe2TiO5)
and rutile (Ti02) being stable above this temperature.