Phase Equilibria in the Fe2O3-P2O 5 System

Article abstract of DOI:10.1111/j.1551-2916.2010.04287.x

Four ferric phosphate compounds were identified in the Fe₂O ₃-P₂O₅ system and the liquidus surfaces in the subsystems Fe₃PO₇-FePO₄, FePO ₄-Fe₄(P₂O₇)₃, and Fe ₄(P₂O₇)₃-Fe(PO₃) ₃ were determined. The results are significantly different from those presented by Wentrup in 1935. Fe₃PO₇ is the stable ferric oxophosphate compound, not Fe₄P₂O₁₁, and Fe₃PO₇ decomposes in air at 1090°C. The congruent melting point of FePO₄ (1208°C) is similar to what was reported, but Fe₄(P₂O₇)₃ melts congruently at 945°C, about 300°C lower than claimed by Wentrup. Fe(PO ₃)₃, for which the melting temperature has not been previously reported, melts congruently at 1205°C. Eutectic points exist at 58.0 mol% Fe₂O₃ (1070°C), 42.7% Fe₂O ₃ (925°C), and 37.0% Fe₂O₃ (907°C). The latter two eutectic points bracket the conventional glass-forming range for iron phosphate melts under consideration as alternative hosts for nuclear wastes.

Full text of DOI:10.1111/j.1551-2916.2010.04287.x

J. Am. Ceram. Soc., 94 [5] 1605–1610 (2011)
DOI: 10.1111/j.1551-2916.2010.04287.x
r 2011 The American Ceramic Society
ournal
J
Phase Equilibria in the Fe2O3–P2O5 System
Liying Zhang,* Mark E. Schlesinger,w and Richard K. Brow*
Department of Materials Science and Engineering, Missouri University of Science and Technology,
Rolla, Missouri 65409
Four ferric phosphate compounds were identified in the Fe2O3–
P2O5 system and the liquidus surfaces in the subsystems Fe3PO7–
FePO4, FePO4–Fe4(P2O7)3, and Fe4(P2O7)3–Fe(PO3)3 were
determined. The results are significantly different from those
presented by Wentrup in 1935. Fe3PO7 is the stable ferric oxo-
phosphate compound, not Fe4P2O11, and Fe3PO7 decomposes in
air at 10901C. The congruent melting point of FePO4 (12081C) is
similar to what was reported, but Fe4(P2O7)3 melts congruently at
9451C, about 3001C lower than claimed by Wentrup. Fe(PO3)3,
for which the melting temperature has not been previously re-
ported, melts congruently at 12051C. Eutectic points exist at 58.0
mol% Fe2O3 (10701C), 42.7% Fe2O3 (9251C), and 37.0%
Fe2O3 (9071C). The latter two eutectic points bracket the con-
ventional glass-forming range for iron phosphate melts under
consideration as alternative hosts for nuclear wastes.
phosphate glass-forming compositions reported in the literature
are centered on the Fe4(P2O7)3–Fe(PO3)2 system15,20,21; however,
no investigation of the liquidus surface of this system has been
reported.
Some information about the phase transition temperatures of
the ferric phosphate compounds has been reported. FePO4 was
reported by Wentrup11 to melt between 12301 and 12401C, and
Shafer et al.22 obtained a melting point of 12301C using rapid
heating in a strip furnace. Caglioti23 mentioned the possible de-
composition of Fe3PO7 to Fe2O3 and FePO4 above 11001C, and
Korinth and Royen12 later confirmed that Fe3PO7 decomposes
at 12001C. Three overlapping endothermic differential thermal
analysis (DTA) peaks between 10001 and 11501C were reported
for Fe3PO7, but were not explained, and the melting tempera-
ture of Fe3PO7 was estimated from DTA to be 13751C.24 Fe(III)
in ferric phosphate systems often undergoes an endothermic
reduction to Fe(II) when heated to the corresponding liquidus
or decomposition temperatures, and P2O5 can volatilize at high
temperatures (generally above 10001C) from melts, particularly
from phosphate-rich compositions. These processes can make the
interpretation of complex thermal curves more difficult, and may
have contributed to the apparent errors in the Wentrup diagram.
Glasses with Fe2O3 contents between 33 and 59 mol% form
from melts held at 13001C,20 and glasses with Fe2O3 contents
between 50 and 63 mol% can form from melts at temperatures
from 11501 to 12501C.25 Such studies provide some information
about the liquidus surface of this system, although it is compli-
cated by the reduction of some ferric ions to ferrous ions under
typical melting conditions. No exact melting temperature for
crystalline Fe(PO3)3 has been reported, although glasses based
on this composition have been prepared from melts quenched
from 12501C.21
I. Introduction
HEMICALLY durable iron phosphate glasses are compatible
C
with a wide variety of other oxides and so have drawn much
attention as alternative hosts for radioactive wastes.1–5 Knowl-
edge of phase equilibria in the iron phosphate system is of
interest for understanding the effects of composition and
temperature on glass formation, and for predicting crystalliza-
tion behavior of iron phosphate melts and glasses.6 In addition,
iron phosphate compounds are finding increasing applications
for use as electrode materials for Li batteries7,8 and for cata-
lysts,9,10 and so the availability of accurate phase equilibrium
information would be useful for the preparation and character-
ization of these materials.
Wentrup determined the original ferric phosphate phase dia-
gram for the subsystems between Fe2O3, Fe4P2O11, FePO4, and
Fe4(P2O7)3 by recording heating and cooling curves of appro-
priate mixtures of Fe2O3, FePO4, and Fe4(P2O7)3.11 One con-
troversial compound in the original Wentrup diagram is the
oxophosphate phase Fe4P2O11. Wentrup reported the formation
of a crystalline phase with this nominal composition, but did not
characterize it. Korinth and Royen12 used X-ray diffraction
(XRD) to study mixtures of Fe2O3 and FePO4 heated at 8001–
9001C and determined that the stoichiometry of the lowest phos-
phate compound was in fact Fe3PO7 instead of Fe4P2O11. Gleitzer
and colleagues studied the solid-state equilibria and formation
of ferric phosphate compounds at 9001C and confirmed that
Fe3PO7, not Fe4P2O11, is the stable oxophosphate compound.13,14
Another controversy associated with the Wentrup phase dia-
gram concerns the FePO4–Fe4(P2O7)3 system. The original phase
diagram shows a melting temperature for ferric pyrophosphate,
Fe4(P2O7)3, above 12001C. However, research on glass formation
in this system indicates that the melting temperature of the
pyrophosphate must be closer to 9501–11001C.15–19 Many iron
In this paper, the existence of four ferric phosphate compounds
is confirmed and the liquidus surface of the ferric phosphate
system is determined between the compounds Fe(PO3)3 and
Fe3PO7. This information is then related to other reported stud-
ies of iron phosphate glass formation and compound formation.
II. Experimental Procedures
(1) Compound Preparation
Ferric phosphate compounds were prepared by solid-state reac-
tions between stoichiometric mixtures of FePO4 and Fe2O3 or
NH4H2PO4 at different temperatures, as summarized in Table I.
FePO4 ꢀ xH2O (100%, Alfa Aesar, Ward Hill, MA), Fe2O3 (Alfa
Aesar, ꢁ 99%), and NH4H2PO4 (Alfa Aesar, 98%) were used
as raw materials. Samples obtained from these experiments were
pulverized too53 mm and characterized by XRD (XRD Scintag
XDS 2000, Scintag Inc., Cupertino, CA) with a slow scanning
rate (11 every 1–2 min). Search and match of XRD patterns were
achieved manually using DMSNT 1.37 (Scintag Inc.), which is
based on a Hanawalt search method.
T. Vanderah—contributing editor
Manuscript No. 28320. Received July 13, 2010; approved October 27, 2010.
This work was financially supported by the National Science Foundation, grants DMR-
0305202 and DMR-0502463.
(2) Phase Equilibria Studies
(A) Fe3PO7–FePO4 System: Samples (o53 mm) weigh-
*Member, The American Ceramic Society.
wAuthor to whom correspondence should be addressed. e-mail: mes@mst.edu
ing 100–200 mg were placed in an open alumina crucible for
1605

1606
Journal of the American Ceramic Society—Zhang et al.
Vol. 94, No. 5
Table I. Preparation Methods for Ferric Phosphate Compounds
Compound
Raw materials
Preparation conditions
Fe3PO7
FePO41Fe2O3
12 h at 9501C, then 72 h at 10501C
Fe4P2O11
FePO4
FePO41Fe2O3, or Fe3PO71FePO4
FePO4 ꢀ xH2O
12–48 h holds at 8001, 9001, 10001, and 10501C—all unsuccessful
12 h at 8801C
12 h at 8001C, then 72 h at 9401C
Ammonia burn-off from thoroughly mixed batch at 5001C overnight,
followed by 12-h hold at 8001C
Fe4(P2O7)3
Fe(PO3)3
FePO41Fe(PO3)3
Fe2O31NH4H2PO4
DTA and thermogravimetric analysis (DTA–TGA, Netzsch
STA 409C/CD, NETZSCH–Geratebau GmbH, Selb, Ger-
¨
III. Results and Discussion
(1) Ferric Phosphate Compounds
many). DTA–TGA was run at 101C/min under flowing air.
The data obtained were analyzed using the Netzsch Proteus soft-
ware, version 4.3. The accuracy of the characteristic temperatures
was determined to be 751C by calibration and multiple runs.
DTA–TGA was used to determine the phase-transition and
decomposition temperatures of crystalline Fe3PO7 and FePO4.
To study the decomposition of Fe3PO7, samples were heated
in air for 10–12 h at temperatures between 10001 and 13501C, and
then quenched in water. These samples were dried, then
either pulverized to o53 mm and analyzed by XRD, or mounted
and coated with carbon for analytical scanning electron micros-
copy with energy-dispersive spectrometry (SEM–EDS, FESEM
S4700, Hitachi High-Tech, Tokyo, Japan). Fe/P ratios were typ-
ically determined by EDS at low magnification ( ꢂ 500) using a
calibration curve based on the EDS analyses of the four ferric
phosphate compounds. In general, analyses were obtained from
at least five different spots on each sample and the average com-
positions are reported. The compositions of the melted regions of
quenched samples obtained by EDS were used to determine the
liquidus surface of the Fe3PO7-rich portion of this system, be-
cause these melts crystallize readily when quenched.25 For the
FePO4-rich portion of this system, similar analyses were per-
formed on samples with the nominal composition (mol%)
51.4Fe2O3–48.6P2O5 heated in air for 12 h at different temper-
atures between 11501 and 12051C.
Single-phase samples of Fe3PO7, FePO4, Fe4(P2O7)3, and
Fe(PO3)3 were prepared according to the processes outlined in
Table I and confirmed by XRD. The compound Fe4P2O11 could
not be obtained under the experimental conditions described in
Table I. Instead, various mixtures of crystalline Fe3PO7 and
FePO4, or Fe2O3 and FePO4, formed. The presence of Fe3PO7
instead of Fe4P2O11 as the stable ‘‘iron-rich’’ oxophosphate
compound in the ferric phosphate system is consistent with
what was reported by Korinth and Royen.12 but in disagreement
with the conclusions of Wentrup.11
(2) Phase Equilibria Studies
(A) Fe3PO7–FePO4 System: Figure 1 shows the XRD
patterns of samples of Fe3PO7 quenched from 10001, 11501, and
12001C after being held in air for 12 h. Crystalline Fe2O3,
FePO4, and Fe2P2O7, as well as a small amount of glassy phase,
are present in the samples heated at or above 11501C. Reactions
1 and 2 summarize the possible decomposition reactions that
account for the formation of these phases:
Fe3PO7 ! Fe2O3 þ Liquid ðFePO4 þ Fe2O3Þ
2 FePO4 ! Fe2P2O7 þ 1=2 O2
(1)
(2)
(B) FePO4–Fe4(P2O7)3 System: One to two grams of
samples with nominal Fe2O3 contents of 44.5 and 47.5 mol%,
quenched after heat treatments similar to those described above,
were prepared and analyzed by SEM–EDS to determine the
compositions of the resulting glassy phase. Fe4(P2O7)3-rich
compositions are good glass formers, so the liquidus surface
was studied by characterizing samples quenched from temper-
atures that bracket the expected liquidus temperature. Samples
about 0.6–1.0 g in size with Fe2O3 contents from 40.0 to 43.0
mol% were sealed under air in silica ampoules to minimize
Fe(III) reduction and P2O5 volatilization during subsequent
thermal treatments. These sealed samples were heated for 12 h
to different temperatures that bracketed the expected liquidus
temperature, followed by a water quench. The temperature
intervals were set at 101C. These quenched samples were ana-
lyzed by optical microscopy (OM) and powder XRD, and the
liquidus surface was determined to be the midpoint between the
highest temperature where crystals were observed and the lowest
temperature where no crystals were observed.
Figure 2 shows DTA and TGA curves for Fe3PO7 and FePO4
heated in air. For Fe3PO7, overlapping endothermic DTA peaks
are present at temperatures around 11001C, consistent with lit-
erature reports about the thermal behavior of Fe3PO7.12,23,24
Several processes, including the decomposition and melting of
Fe3PO7 (reaction (1)) and the reduction of Fe(III) (reaction (2)),
may account for these endothermic events. The reduction of
Fe(III) to Fe(II) accounts for the TGA weight loss. EDS ana-
lyses indicate that the overall Fe/P ratio of an Fe3PO7 sample
heat treated at 12001C for 12 h was 3.0070.16, consistent with
(c)
1200°C
(b)
1150°C
(C) Fe4(P2O7)3–Fe(PO3)3 System: P2O5-rich melts are
good glass formers and so the P2O5-rich portion of the diagram
was also studied by analyzing samples quenched from temper-
atures that bracket the expected liquidus temperatures. These
samples (compositional intervals of B1.5 mol% Fe2O3) were
sealed in silica ampoules and heated in a similar way to what is
described above for the Fe4(P2O7)3-rich samples. The quenched
samples were studied by OM and micro-Raman spectroscopy
(Horiba–Jobin Yvon LabRam-HR, Horiba Jobin Yvon Inc.,
Edison, NJ) using a He–Ne laser (632.8 nm). Raman spectra of
crystalline Fe4(P2O7)3 and Fe(PO3)3 are more distinct than the
respective XRD patterns, and so this technique was convenient
for identifying isolated crystals in these quenched samples.
(a)
1000°C
20
25
30
35
40
45
50
55
60
2 Theta (°)
Fig. 1. X-ray diffraction patterns of Fe3PO7 quenched after 12 h in air
from (a) 10001C, (b) 11501C, and (c) 12001C. &, Fe3PO7 (JCPDS: 37-
0061); ., FePO4 (JCPDS: 84-0875); ,, Fe2O3 (JCPDS: 33-0664), and
J, Fe2P2O7 (JCPDS: 72-1516).

May 2011
Phase Equilibria in the Fe2O3–P2O5 System
1607
100
FePO4
99
Fe3PO7
98
97
96
95
Solid phase
transition
FePO4
β→γ
α→β
200
400
600
800
1000
1200
Temperature (°C)
Fig. 2. Differential thermal analysis and thermogravimetric analysis
patterns for Fe3PO7 and FePO4.
the initial stoichiometry of the sample, and indicating that no
significant loss of P2O5 occurred from these iron-rich melts.
Based on the results shown in Figs. 1 and 2, the decomposition
temperature of Fe3PO7 is estimated to be 10901781C.
The DTA data for Fe3PO7 in Fig. 2 reveal a solid-state phase
transition between 8401 and 8801C. This phase transition could
be detected by DTA on cooling as well (not shown). Wentrup11
reported a solid-state phase transition temperature of 8691C for
his oxophosphate phase (Fe4P2O11). The identification of a solid-
state phase transition at a similar temperature in Fe3PO7 is taken
as evidence that Wentrup misidentified Fe3PO7 as Fe4P2O11. The
DTA data for FePO4 in Fig. 2 indicates that there are two solid-
state phase transitions, at 7101751C (a-b) and at 8801751C
(b-g). These transition temperatures are similar to those previ-
ously reported (7071 and 8891C, respectively).11,22
Fig. 3. Backscattered electron images of Fe3PO7 quenched after dwell-
ing for 12 h at 12001C; (a) and (b) are from the same sample but at
different magnifications.
Figure 3 shows backscattered electron images of an Fe3PO7
sample quenched from 12001C after being held there in air for
12 h. EDS analyses reveal that the bright sphere-like regions are
free of phosphorus and appear to be Fe2O3. EDS and XRD in-
dicate that the small crystals (o2 mm) formed around the spher-
ical particles include FePO4, Fe2P2O7, and Fe2O3, the phases
expected to precipitate from the liquid phase that forms at this
temperature (reactions (1) and (2)). The average composition of
the regions around the large Fe2O3 particles was determined by
EDS and is used as the composition of the liquid phase at the
respective heat-treatment temperatures. Similar analyses were
performed on other compositions in this system, and these ex-
perimental data are used for the liquidus surface for the Fe3PO7–
FePO4 system that is plotted in Fig. 4; error bars indicate
uncertainties in the quantitative EDS analyses. By extrapolating
the exponential line fitting these experimental points, the tem-
peratures at which Fe3PO7 and Fe2O3 are predicted to fully melt
are estimated to be 13801 and 16001C, respectively. These tem-
peratures are similar to those reported in the literature (13751C
portion of the FePO4–Fe4(P2O7)3 system. For the Fe4(P2O7)3-
rich part of this system, samples were sealed in silica ampoules
to minimize the effects of phosphorus volatility at high temper-
atures. EDS analyses of a sample of Fe4(P2O7)3 heated at 9501C
for B12 h in air and in a sealed ampoule indicate Fe/P ratios of
0.7570.05 and 0.7070.02, respectively, compared with an ex-
pected ratio of 0.67. Figure 5 shows the XRD patterns collected
from several compositions quenched from temperatures below
their respective liquidus temperatures. These results were used to
determine the liquidus temperatures and the eutectic point of the
FePO4–Fe4(P2O7)3 system.
Figure 6 shows the liquidus curves obtained by fitting the
points obtained from the analyses described above. The FePO4–
1600
1500
1400
1300
1200
1100
1000
900
24
for Fe3PO7 and 15651C for Fe2O3 26).
The FePO4–Fe3PO7 system has a eutectic point at 58.071.2
mol% Fe2O3, and the eutectic reaction occurs at 10701751C.
The eutectic composition and temperature of the FePO4–
Fe4P2O11 system reported by Wentrup11 were 58.0 mol%
Fe2O3 and 9681C. Wentrup appears to have misidentified
Fe3PO7 as Fe4P2O11, and may have misinterpreted endothermic
evidence of Fe(III) reduction for eutectic melting. The eutectic
temperature reported here was confirmed by observing the melt-
ing behavior of the eutectic composition. Glasses can be formed
from melts with compositions around this eutectic point using
rapid quench techniques, as reported elsewhere.25
800
700
600
FePO4
55
65
Fe3PO7
85
95 Fe2O3
Fe2O3 mole % →
(B) FePO4–Fe4(P2O7)3 System: The same sample prep-
aration and characterization methods described above were
used to determine the liquidus surface of the FePO4-rich
Fig. 4. Phase diagram of the FePO4–Fe3PO7 system. Dashed-dotted line
is an extrapolation of the curve used to fit the experimental liquidus points.

1608
Journal of the American Ceramic Society—Zhang et al.
Vol. 94, No. 5
200 400 600 800 1000 1200 1400 1600 1800 2000
Raman shift (cm–1
20
30
40
50
60
)
2 Theta (°)
Fig. 7. Raman spectra of several partially crystallized samples
quenched from the indicated temperatures and from crystalline
Fe(PO3)3 and Fe4(P2O7)3.
Fig. 5. X-ray diffraction patterns of some compositions quenched from
the indicated subliquidus temperature, compared with crystalline FePO4
and Fe4(P2O7)3.
Fe4(P2O7)3 system has a eutectic point at 42.770.4 mol% Fe2O3
and 9251781C. In contrast, Wentrup11 reported a liquidus point
at 46.5 mol% Fe2O3 and 9541C.
rizes the characteristic temperatures determined for this system.
Four ferric phosphate compounds appear in this liquidus sur-
face phase diagram; Fe3PO7 decomposes during heating, and
the others melt congruently. Eutectic points exist in each sub-
system. Below the corresponding liquidus surface, neighboring
solid compounds in the diagram are expected to coexist at equi-
librium, although the details of the solid-phase equilibria were
not studied in this work.
(C) Fe4(P2O7)3–Fe(PO3)3 System: EDS analyses of a
sample of Fe(PO3)3 heated in a silica ampoule to 12501C for 4 h
showed an Fe/P ratio of 0.3770.07, compared with an expected
ratio of 0.33. The silica content in this sample was o2.0 mol%,
as measured at a distance of 1 mm from the ampoule wall on a
sample that was 11 mm in diameter. This analysis indicates that
the nominal composition of this P2O5-rich sample was retained
when melted in an ampoule.
Figure 7 shows the Raman spectra collected from crystals
detected in samples from the Fe4(P2O7)3–Fe(PO3)3 system
quenched from temperatures below their respective liquidus tem-
peratures. By comparing the Raman spectra of these samples to
those collected from crystalline Fe(PO3)3 and Fe4(P2O7)3, the
equilibrium crystal phases at the quenching temperatures could
be determined. A detailed discussion of the Raman spectra
from iron phosphate compounds and glasses will be reported
elsewhere.
This work confirms the reports of Korinth and Royen12 and
Gleitzer and colleagues13,14 that Fe3PO7 (3Fe2O3–P2O5) is the
only ferric oxophosphate phase with an iron content between
FePO4 and Fe2O3. The solid-state phase-transition temperature,
decomposition temperature, and liquidus temperature deter-
mined in the present work are in good agreement with the re-
spective temperatures reported in these earlier studies. The ferric
oxophosphate phase reported in the Wentrup phase diagram,11
Fe4P2O11 (2Fe2O3–P2O5), could not be produced. This finding
has implications for the development of ferric oxophosphate
compounds of interest for electrode materials. For example, the
formation of Fe3PO7 as an electrode material for Li batteries
was accomplished by heating stoichiometric mixtures of Fe2O3
and FePO4 at 10501C for 12 h.7 This temperature is just below the
eutectic temperature (10701C) of the Fe3PO7–FePO4 system
determined in this study (Fig. 4).
Figure 8 shows the liquidus surface for the Fe4(P2O7)3–
Fe(PO3)3 system obtained from the OM–Raman and EDS ana-
lyses. The eutectic point is 37.070.3 mol% Fe2O3 with a eutectic
temperature of 9071781C. This appears to be the first report of
the liquidus surface and eutectic composition in this subsystem.
Glass formation around the eutectic composition in the
FePO4–Fe3PO7 system has been evaluated and is described else-
where.25 The viscosities of those ferric phosphate melts near their
respective liquidus temperatures were relatively low, necessitating
the use of rapid quenching techniques (41031C/s) to form
glasses. These quench rates were much faster than those needed
(3) Discussion of the Liquidus Surface Determination
Figure 9 summarizes the liquidus surface of the ferric phosphate
system between 25 and 75 mol% Fe2O3, and Table II summa-
1400
Wentrup11
This work
1300
1200
Liquid
1200
Liquid
1100
1100
L+ Fe(PO3)3
1000
L + FePO4
1000
L + Fe4(P2O7)3
964°C
900
907°C
900
925°C
Fe(PO3)3+Fe4(P2O7)3
Fe4(P2O7)3 + FePO4
800
800
Fe4(P2O7)3
42
44
46
48
FePO4
28
30
32
34
36
38
Fe2O3 mole% →
Fe2O3 mole% →
Fig. 6. Liquidus surface of the FePO4–Fe4(P2O7)3 system.
Fig. 8. Liquidus surface of the Fe4(P2O7)3–Fe(PO3)3 system.

May 2011
Phase Equilibria in the Fe2O3–P2O5 System
1609
phate compound that forms and it decomposes in air at 10901C.
The liquidus temperature information in the FePO4–Fe3PO7
subsystem is consistent with literature reports on processing
materials of interest for electrodes in lithium electrochemical
devices. Fe4(P2O7)3 was found to melt congruently at 9451C,
about 3001C lower than in earlier claims. For the first time, the
liquidus surface of the Fe(PO3)3–Fe4(P2O7)3 subsystem has been
reported. Fe(PO3)3 melts congruently at 12051C and a eutectic
point exists at 37.0% Fe2O3 (9071C). The liquidus temperatures
of this subsystem are consistent with glass formation and crys-
tallization behavior of compositions being developed for waste
vitrification applications.
1600
1500
1400
1300
1200
1100
1000
900
800
700
600
Fe(PO3)3
40
50
60
70
80
90
Fe2O3
Fe2O3 mole% →
Fig. 9. Summary of the liquidus surface of the ferric phosphate system.
Acknowledgments
The authors are grateful to Dr. Eric W. Bohannan, Clarissa Wisner, and Jong
Wook Lim, Missouri University of Science and Technology, for their assistance
with the XRD, SEM–EDS, and Raman analyses, respectively. Fruitful discussions
with Dr. Luciana Ghussn and Prof. Edgar Zanotto (Sao Carlos University, Sao
˜ ˜
Carlos, Brazil) on the crystallization behavior of iron phosphate melts are greatly
appreciated.
Table II. Summary of the Characteristic Temperatures in the
Fe3PO7–Fe(PO3)3 System
Fe2O3
Temperature
(mol%)
(1C)
Description
75.0
109078 Decomposition of Fe3PO7
References
1M. G. Mesko and D. E. Day, ‘‘Immobilization of Spent Nuclear Fuel in Iron
Phosphate Glass,’’ J. Nucl. Mater., 273, 27–36 (1999).
58.071.2 107075 Eutectic melting (FePO4–Fe3PO7)
50.0
42.770.4
40.0
120878 Congruent melting of FePO4
92578 Eutectic melting (Fe4(P2O7)3–FePO4)
94578 Congruent melting of Fe4(P2O7)3
90778 Eutectic melting (Fe(PO3)3–
Fe4(P2O7)3)
2X. Yu, G. J. Long, R. K. Brow, and D. E. Day, ‘‘Properties and Structure of
Sodium–Iron Phosphate Glasses,’’ J. Non-Cryst. Solids, 215, 21–31 (1997).
3W. Huang, D. E. Day, C. S. Ray, and C. W. Kim, ‘‘High Temperature Prop-
erties of an Iron Phosphate Melt Containing High Chrome Nuclear Waste,’’
J. Nucl. Mater., 346, 298–305 (2005).
37.070.3
4C. W. Kim, C. S. Ray, D. Zhu, D. E. Day, D. Gombert, A. Aloy,
´
A. Mogug-Milankovic, and M. Karabulut, ‘‘Chemically Durable Iron Phosphate
33.3
120578 Congruent melting of Fe(PO3)3
Glasses for Vitrifying Sodium Bearing Waste (SBW) Using Conventional and
Cold Crucible Induction Melting (CCIM) Techniques,’’ J. Nucl. Mater., 322,
152–64 (2003).
to form glasses with compositions in the Fe(PO3)3–Fe4(P2O7)3
system. As shown in Fig. 9, the liquidus temperatures in the
compositional range of approximately 36–43 mol% Fe2O3 are
below 9501C. This is the compositional range of greatest interest
for vitrifying nuclear wastes17–19 and the relatively low liquidus
temperatures determined in this study are consistent with the
glass-forming tendencies reported in the literature.
5M. Karabulut, G. K. Marasinghe, C. S. Ray, D. E. Day, G. D. Waddill, C. H.
Booth, P. G. Allen, J. J. Bucher, D. L. Caulder, and D. K. Shuh, ‘‘An Investi-
gation of the Local Iron Environment in Iron Phosphate Glasses Having
Different Fe(II) Concentrations,’’ J. Non-Cryst. Solids, 306, 182–92 (2002).
6A. Mogug-Milankovic
, M. Rajic, A. Dragner, R. Trojko, and D. E. Day,
´ ´
‘‘Crystallization of Iron Phosphate Glasses,’’ Phys. Chem. Glasses, 39, 70–5 (1998).
7Y.-S. Hong, Y. J. Park, K. S. Ryu, and S.H Chang, ‘‘Crystalline Fe3PO7
as an Electrode Material for Lithium Secondary Batteries,’’ Solid State Ionics.,
156, 27–33 (2003).
To the best of our knowledge, this is the first report of liq-
uidus temperature information for ferric phosphate melts in the
Fe4(P2O7)3–Fe(PO3)3 system. Part of the difficulty in obtaining
accurate information about these materials is avoiding the va-
porization of P2O5, particularly at the higher temperatures re-
quired as compositions approach Fe(PO3)3. Glass formation
from phosphate-rich melts has been reported,21 but those glasses
appear to have lost some phosphate and were subject to some
iron reduction during the melting process.
Efforts were made in this study to minimize the effects of
phosphate volatility and the reduction of ferric phases, but the
compositional uncertainties associated with both processes may
affect these final results. The compositional dependence of the
liquidus temperatures of the ferric phosphate melts shown in
Fig. 9 can be used for guidance in understanding the behavior of
iron phosphate melts. However, these melts will reduce in air at
typical melting temperatures (10001–13001C), to produce glasses
with about 20% ferrous ions.15 Information about the liquidus
temperatures of iron phosphate glass-forming melts is limited,
but DTA studies of crystallization behavior of iron phosphate
glasses reveal crystal melting temperatures near 9001C,6,19,27
consistent with the liquidus temperature for the ferric phosphate
compositions summarized in Fig. 9.
8C. M. Burba, J. M. Palmer, and B. S. Holinsworth, ‘‘Laser-Induced Phase
Changes in Olivine FePO4: A Warning on Characterizing LiFePO4-Based Cath-
odes with Raman Spectroscopy,’’ J. Raman Spectrosc., 40, 225–8 (2009).
9M. Ai and K. Ohdan, ‘‘Effects of the Method of Preparing Iron Orthophos-
phate Catalyst on the Structure and the Catalytic Activity,’’ Appl. Catal., A, 180,
47–52 (1999).
10Y. Wang, X. X. Wang, Z. Su, Q. Guo, Q. Tang, Q. Zhang, and H. Wan,
‘‘SBA-15-Supported Iron Phosphate Catalyst for Partial Oxidation of Methane to
Formaldehyde,’’ Catal. Today, 155, 93–5 (2004).
11H. Wentrup, ‘‘Contribution on the System Iron–Phosphorus–Oxygen,’’ Arch.
Eisenhut., 9, 57–60 (1935).
12J. Korinth and P. Royen, ‘‘Reaction in the System Fe2O3/FePO4,’’ Z. Anorg.
Allg. Chem., 313, 121–37 (1961).
¨
13J. C. Kaell, F. Jeannot, and C. Gleitzer, ‘‘Study of the Progressive Reduction
of Fe3(PO4)2 and Fe9(PO4)O8,’’ Ann. Chim., 9, 169–80 (1984).
14A. Modaressi, J. C. Kaell, B. Malaman, R. Ge
´
rardin, and C. Gleitzer, ‘‘Study
of the System Fe–P–O (for Fe/P ꢁ 1) and Its Compounds: The Oxyphosphates of
Iron,’’ Mater. Res. Bull., 18, 101–9 (1983).
15X. Fang, C. S. Ray, A. Mogus-Milankovic, and D. E. Day, ‘‘Iron Redox
Equilibrium, Structure and Properties of Iron Phosphate Glasses,’’ J. Non-Cryst.
Solids, 283, 162–72 (2001).
16S. T. Reis, M. Karabulut, and D. E. Day, ‘‘Structural Features and Properties
of Lead–Iron–Phosphate Nuclear Wasteforms,’’ J. Nucl. Mater., 304, 87–95 (2002).
17D. E. Day, Z. Wu, C. S. Ray, and P. Hrma, ‘‘Chemically Durable Iron Phos-
phate Glass Wasteforms,’’ J. Non-Cryst. Solids, 241, 1–12 (1998).
18C. W. Kim and D. E. Day, ‘‘Immobilization of Hanford LAW in Iron Phos-
phate Glasses,’’ J. Non-Cryst. Solids, 331, 20–31 (2003).
19R. J. Bingham, P. A. Hand, S. D. Forder, A. Lavaysierre, S. H. Kilcoyne, and
I. Yasin, ‘‘Preliminary Studies of Sulphate Solubility and Redox in 60P2O5–
40Fe2O3 Glasses,’’ Mater. Lett., 60, 844–7 (2006).
20G. K. Marasinghe, M. Karabulut, C. S. Ray, D. E. Day, M. G. Shumsky,
W. B. Yelon, C. H. Booth, P. G. Allen, and D. K. Shuh, ‘‘Structural Features of
Iron Phosphate Glasses,’’ J. Non-Cryst. Solids, 222, 144–52 (1997).
VI. Conclusions
21B. Pivac, A. Mogug-Milankovic
´
, and D. E. Day, ‘‘Iron Valence and Coordi-
The liquidus surface of the Fe3PO7–Fe(PO3)3 system was re-
determined using sample preparation techniques that minimize
the effects of sample volatilization and reduction. Significant
differences are found when this work is compared with the ear-
lier reports. In particular, Fe3PO7 is the only ferric oxophos-
nation in Phosphate Glasses as Studied by Optical Spectroscopy,’’ J. Non-Cryst.
Solids, 226, 41–6 (1998).
22E. C. Shafer, M. W. Shafer, and R. Roy, ‘‘Studies of Silica Structure Phases:
II. Data on FePO4, FeAsO4, MnPO4, BPO4, AIVO4, and Others,’’ Z. Kristallogr.,
108, 263–75 (1956).

1610
Journal of the American Ceramic Society—Zhang et al.
Vol. 94, No. 5
23V. Caglioti, ‘‘The Structure of Ferric Phosphate,’’ Atti accad. Lincei., 22,
26D. L. Perry and S. L. Phillips, Handbook of Inorganic Compounds. CRC Press,
Boca Raton, (1995).
146–9 (1935).
24M. R. De Guire, T. R. S. Prasanna, G. Kalonji, and R. C. O’Handley, ‘‘Phase
Equilibria in the Iron Oxide–Cobalt Oxide–Phosphorus Oxide System,’’ J. Am.
Ceram. Soc., 70, 831–7 (1987).
27L. Zhang, L. Ghussn, M. L. Schmitt, E. D. Zanotto, R. K. Brow, and M. E.
Schlesinger, ‘‘Thermal Stability of Glasses from the Fe4(P2O7)3–Fe(PO3)3 System,’’
J. Non-Cryst. Solids (2010) (in press).
&
25L. Zhang, R. K. Brow, M. E. Schlesinger, L. Ghussn, and E. D. Zanotto,
‘‘Glass Formation from Iron-Rich Phosphate Melts,’’ J. Non-Cryst. Solids, 356,
252–7 (2010).