A Raman study of iron-phosphate crystalline compounds and glasses

Article abstract of DOI:10.1111/j.1551-2916.2011.04486.x

Ferrous and ferric phosphate crystalline compounds and glasses were studied using Raman spectroscopy. A comparison of the spectra from crystalline and glassy ortho-, pyro-, and metaphosphates indicates that similar phosphate anions constitute the structures of the respective materials, and some information about the compositional dependence of the phosphate-site distributions in the glasses can be gleaned from relative peak intensities. A correlation exists between the average P-O bond distance and the Raman peak frequencies in the crystalline compounds, and this correlation is used to provide information about the structures of the iron phosphate glasses. For example, the average P-O bond distance is estimated to decrease from about 1.57 A for iron metaphosphate glasses (O/P～3.0) to 1.54 A for iron orthophosphate glasses (O/P～4.0). These bond distances are in good agreement with those reported from diffraction studies of similar glasses.

Full text of DOI:10.1111/j.1551-2916.2011.04486.x

J. Am. Ceram. Soc., 94 [9] 3123–3130 (2011)
DOI: 10.1111/j.1551-2916.2011.04486.x
r 2011 The American Ceramic Society
ournal
J
A Raman Study of Iron–Phosphate Crystalline Compounds and Glasses
Liying Zhang* and Richard K. Brow*,w
Department of Materials Science and Engineering, Missouri University of Science and Technology, Rolla 65409, Missouri
Ferrous and ferric phosphate crystalline compounds and glasses
were studied using Raman spectroscopy. A comparison of the
spectra from crystalline and glassy ortho-, pyro-, and metaphos-
phates indicates that similar phosphate anions constitute the struc-
tures of the respective materials, and some information about the
compositional dependence of the phosphate-site distributions in the
glasses can be gleaned from relative peak intensities. A correlation
exists between the average P–O bond distance and the Raman
peak frequencies in the crystalline compounds, and this correlation
is used to provide information about the structures of the iron
phosphate glasses. For example, the average P–O bond distance is
estimated to decrease from about 1.57 A for iron metaphosphate
glasses (O/PB3.0) to 1.54 A for iron orthophosphate glasses
(O/PB4.0). These bond distances are in good agreement with
those reported from diffraction studies of similar glasses.
glasses and crystals.11–34 The frequency of the P–O-stretching
vibrations changes systematically with the number of bridging
oxygens (Qx) on a tetrahedron, and so Raman peaks associated
with the different P–O vibrational modes can be used to identify
different structural elements. Rulmont et al.15 compared the Ra-
man and IR spectra of pyro- and metaphosphate materials and
showed that crystalline and glassy phosphates with similar com-
positions have similar phosphate anions. In addition, more quan-
titative information about the structures of crystalline and glassy
phosphates, including estimates of P–O bond lengths and P–O–P
bond angles, can be obtained from Raman peak positions.15,16
There have been several Raman studies of iron phosphate
glasses.17–19 The compositions studied were generally limited to
those near the pyrophosphate stoichiometry of interest for waste
vitrification applications. Qualitative changes in peak shapes
and positions have been related to glass compositions, but little
detailed information has been reported. In the present study, the
Raman spectra of 10 crystalline ferric, ferrous, and mixed ferric-
ferrous phosphate compounds, including ortho- (O/P5 4.0, Q0),
pyro- (O/P5 3.5, Q1), and metaphosphate (O/P5 3.0, Q2) com-
pounds, were collected, and those results are used to interpret
the Raman spectra of iron phosphate glasses with a wide variety
of O/P and Fe/P ratios to provide information about phosphate
tetrahedral distributions and estimates of P–O bond lengths for
a much broader range of iron phosphate glass compositions
than has been reported previously.
I. Introduction
RYSTALLINE and amorphous iron phosphate materials are
being developed for a variety of technological applications.
C
For example, amorphous and crystalline FePO4 and similar
compounds have been developed as catalysts,1 and the catalytic
performance is affected by reduction to Fe2P2O7.2 LiFePO4 has
been proposed as a cathode material for rechargeable Li-ion
batteries3,4 and the dilithiation process can form disordered prod-
ucts that are sometimes difficult to characterize by conventional
diffraction techniques. The ferric oxo-phosphate phase Fe3PO7
has also been evaluated as a potential electrode material.5
Iron phosphate glasses are of interest for a variety of appli-
cations, including as corrosion resistant hosts for radioactive
wastes.6–9 Typical iron phosphate glasses for waste applications
are based on a ferric pyrophosphate (40Fe2O3–60P2O5 molar)
nominal composition in which some fraction of ferric ions
reduce to ferrous ions, to yield a structure based on ferric and
ferrous polyhedra that link various phosphate anions.8,9 The
properties of these glasses are sensitive to changes in iron va-
lence and the Fe/P ratio, both of which affect the overall O/P
ratio, which determines the distribution of phosphate anions.
Glasses with O/PB3 are classified as metaphosphates and pos-
sess relatively long chains of P-tetrahedra that link neighboring
tetrahedra through two bridging oxygens; these tetrahedra are
sometimes classified as Q2-tetrahedra.10 The chains are termi-
nated by phosphate units with a single bridging oxygen (Q1
units). A pyrophosphate composition (O/PB3.5) could have a
structure based principally on Q1-tetrahedra that form P2O47ꢀ
anions. Compositions with O/P43.5 will have structures based
on Q1- and Q0-tetrahedra; the latter are similar to isolated (no
bridging oxygens) phosphate tetrahedra found in crystalline or-
thophosphate compounds.
II. Experimental Procedures
Raw materials including stoichiometric FePO4 ꢁ xH2O (100%,
Alfa Aesar), Fe2O3 (Alfa Aesar, ꢂ99%), and NH4H2PO4 (Alfa
Aesar, 98%, Ward Hill, MA) were used to prepare the crystal-
line compounds and glasses. Iron phosphate crystalline com-
pounds were prepared following the procedures listed in Table I.
Additional details on the preparation and properties of ferric
phosphate compounds can be found in Zhang et al.35 X-ray
diffraction (Scintag XDS 2000, Scintag, Inc., Cupertino, CA)
was used to confirm that the desired, stoichiometic phases were
formed. Secondary crystalline phases could not be detected by
XRD in any of these samples.
Glasses were prepared from iron phosphate crystalline com-
pounds or batch materials. The melt conditions are summarized
in Table II. Some melts with O/P ratios ꢂ 4.0 were quenched
between steel rollers to avoid crystallization.19 Other melts with
O/P ratios near 3.0 were prepared in sealed silica ampoules to
minimize P2O5-volatilization and to control the iron valence
during melting. The ampoules were quenched in water after the
designated melt time. Finally, several melts with intermediate
[Fe]/[P] ratios were prepared by conventional methods, in air in
open alumina crucibles and then quenched on steel plates. In
general, the sample sizes were 3–5 g for orthophosphate glasses
prepared by the roller quenching method, 0.6–2.5 g for glasses
melted in sealed silica ampoules, and 20–40 g for the glasses
melted in open crucibles. Every glass was pulverized to B53 mm
and characterized by XRD to confirm the vitreous state.
Samples of glass powders were coated by carbon and their Fe/P
ratios were determined using the energy-dispersive X-ray spect-
rometry (EDS) system associated with the Hitachi S4700 scanning
electron microscope (Hitachi High-Tech, Tokyo, Japan). These
Raman spectroscopy has been widely used to provide informa-
tion about the anions that constitute the structures of phosphate
L. Pinckney—contributing editor
Manuscript No. 28759. Received October 11, 2010; approved February 2, 2011.
This work was supported by the National Science Foundation under Grant No. DMR-
0305202 and DMR-0502463.
*Member, The American Ceramic Society.
wAuthor to whom correspondence should be addressed. e-mail: brow@mst.edu
3123

3124
Journal of the American Ceramic Society—Zhang and Brow
Vol. 94, No. 9
Table I. Conditions Used to Prepare the Iron Phosphate Crystalline Compounds
Compound
Batch materials
Preparation conditions
a-FePO4
FePO4 ꢁ xH2O
8801C in air for 12 h
Reduced in forming gas at 680–6901C for 6 h
11501C for 12 h in sealed ampoule
9001C for 12 h in sealed ampoule
8001C in air for 12 h, then 9401C in air for 24–48 h
Ã
Fe3(PO4)2-A
Fe3(PO4)2-B
Fe7(PO4)6
Fe4(P2O7)3
Fe2P2O7
FePO4, Fe2O3
Fe3(PO4)2-A
Fe2O3, FeP2O6
FePO4, Fe(PO3)3
FePO4
Ã
Reduced in forming gas at 5601C for 6 h
Fe3(P2O7)2
Fe7(P2O7)4
Fe(PO3)3
Fe(PO3)3, Fe2PO5
Fe2P2O7, Fe3(P2O7)2
Fe2O3, NH4H2PO4
FePO4, NH4H2PO4
9001C for 12 h in sealed ampoule
9001C for 12 h in sealed ampoule
Ammonia burn-off at 5001C overnight, held in air at 8001C for 12 h
Ammonia burn-off at 5001C overnight, then reduced in forming gas
at 6501C for 6 h
Ã
Fe(PO3)2
Ã
Forming gas is 10% H2 and 90% Ar.
analyses were based on an Fe/P calibration curve determined by
analyzing the corresponding crystalline compounds. At least five
measurements were done on each sample, and the average Fe/P
ratio, with one standard deviation, is reported. The Fe21/Fetot
contents of the glasses were determined by a titration technique
using KMnO4 (B2 mM),36 with an absolute uncertainty of 2%.
The Fe/P and Fe21/Fetot ratios were then used to calculate the O/
P ratio for every glass. EDS analyses indicated that these glasses
possessed o2 mol% Al2O3 (or SiO2) as a contaminant from the
crucible (or ampoule). There was no evidence for these contam-
inants affecting the Raman spectra of the glasses, and their pres-
ence is ignored in discussions of glass compositions and
structures.
An Horiba–Jobin Yvon LabRam-HR spectrometer (Horiba-
Jobin Yvon, Inc., Edison, NJ) was used to collect Raman spec-
tra with a He–Ne laser (632.8 nm) as the excitation source. In
general, spectra were collected through a ꢃ 10 microscope ob-
jective from the surfaces of crystalline powders (53 mm particle
size) and from the surfaces of fragments of glasses prepared by
the various methods.
ratios obtained by EDS. In general, the O/P ratios differ from
their nominal values principally because of a change in the
average Fe-redox state after melting.
Figure 1 shows the Raman spectra collected from the crys-
talline iron orthophosphate (O/P5 4) compounds. The major
band near B1009 cmꢀ1 in the spectrum from a-FePO4 is
assigned to the symmetric PO4-stretching mode associated
with the Q0 PO34ꢀ tetrahedra.20,21 For Fe3(PO4)2-A and -B, the
intense bands between 900 and 980 cmꢀ1 are also assigned to
the symmetric PO4-stretching modes. The less intense bands
between 900 and 1100 cmꢀ1 in the spectra from Fe7(PO4)6,
Fe3(PO4)2-A, and Fe3(PO4)2-B are assigned to the asymmetric
PO4 modes associated with the reduced symmetry of these PO43ꢀ
units.22 The peaks below 600 cmꢀ1 are related to different P–O
and Fe–O stretching and bending modes.21,37
Figure 1 also shows the Raman spectra collected from the
glasses that have nominal compositions near the iron ortho-
phosphate stoichiometry (O/Pꢄ4). The Raman spectra from
the O1, O2, and O3 glasses are similar to that obtained from
a-FePO4. The spectra from these glasses are dominated by an
intense peak centered between 1000 and 1010 cmꢀ1, due to the
PO4-stretching modes associated with the nonbridging oxygens,
and some lower intensity peaks are also present in the range
between 200 and 500 cmꢀ1, similar to what is seen for a-FePO4.
The broader full-widths at half-maximum (FWHM 5 30–40
cmꢀ1) of the peak near 1000 cmꢀ1 and their slightly lower fre-
quencies from the glasses, compared with crystalline FePO4
(FWHM 5 10 cmꢀ1), are consistent with what was reported
by Burba et al.20 for amorphous and crystalline a-FePO4, but
differ from those spectral features reported for the orthorhombic
form of FePO4. The latter spectra resemble that shown in Fig. 1
for Fe3(PO4)2-A. The assignments for the low-frequency peaks
in the spectra from the O1, O2, and O3 glasses are the same as
those described above for the crystalline samples.
III. Results
Table III lists the crystallographic parameters reported in the
literature for the iron phosphate crystals prepared in this study.
The average P–O bond distances for nonbridging (P–Onb) and
bridging (P–Obr) oxygens are indicated. Nonbridging oxygens
are those that are linked to one P-tetrahedron in the structure,
and bridging oxygens are linked to two P-tetrahedra. Also listed
are the average P–O–P bond angles for the crystalline pyro- and
metaphosphate compounds.
Table IV summarizes the compositions of the glasses pre-
pared in this work. The O/P ratios were calculated from the
measured Fe21/Fetotal ratios, obtained by titration, and the Fe/P
Table II. Conditions Used to Prepare the Iron Phosphate Glasses
Glass
Nominal composition (mole fraction)
Melting vessel
Atmosphere
Temp (1C)/ time (h)
Quench methods
O1
O2
O3
P1
P2
P3
P4
P5
P6
P7
60Fe2O3–40P2O5
52Fe2O3–48P2O5
50Fe2O3–50P2O5
40Fe2O3–60P2O5
67FeO–33P2O5
Al2O3 crucible
Al2O3 crucible
Al2O3 crucible
Al2O3 crucible
SiO2 ampoule
SiO2 ampoule
Al2O3 crucible
SiO2 ampoule
Al2O3 crucible
Al2O3 crucible
Al2O3 crucible
SiO2 ampoule
SiO2 ampoule
SiO2 ampoule
SiO2 ampoule
In air
In air
In air
In air
Sealed under vacuum
Sealed under vacuum
In air
Sealed in air
In air
In air
1200/0.5 h
1300/0.5 h
1300/2 h
1100/2 h
1200/2 h
1200/3 h
1200/2 h
1110/12 h
1200/2 h
1150/2 h
1150/2 h
1180/12 h
1250/4 h
1200/2 h
1200/0.5 h
Roller quench19
Roller quench19
Roller quench19
Steel plate quench
Water quench
Water quench
Steel plate quench
Water quench
Steel plate quench
Steel plate quench
Steel plate quench
Water quench
Water quench
Water quench
Water quench
10Fe2O3–50FeO–40P2O5
35Fe2O3–65P2O5
31Fe2O3–69P2O5
45Fe2O3–55P2O5
45Fe2O3–55P2O5
41Fe2O3–59P2O5
28Fe2O3–72P2O5
25Fe2O3–75P2O5
50FeO–50P2O5
P8
In air
M1
M2
M3
M4
Sealed in air
Sealed in air
Sealed under vacuum
Sealed under vacuum
50FeO–50P2O5

September 2011
Raman Study of Iron Phosphates
3125
Table III. Crystallographic Parameters Reported for the Crystalline Iron Phosphates
Compound
Space group
Average P–Onb (A)w
Average P–Obr (A)w
P–O–P bond angle
References
FePO4
P3121
P21/c
P21/c
P 1
P21/n
ꢀ
C 1
Pnma
C2221
Ic
1.530
1.537
1.540
1.543
1.514
1.519
1.506
1.512
1.484
1.485
—
—
—
—
1.575
1.554
1.593
1.596
1.577
1.59
—
—
—
—
Goiffon et al.45
Warner et al.46
Warner et al.47
Fe3(PO4)2-A
Fe3(PO4)2-B
Fe7(PO4)6
Fe4(P2O7)3
Fe2P2O7
Fe3(P2O7)2
Fe7(P2O7)4
Fe(PO3)3
Gorbunov et al.48
Elbouaanani et al.49
Hoggins et al.50
Ijjaali et al.51
ꢀ
155.71
—
z
135.21
136.51
1431y
Malaman et al.52
Rojo et al.53
Fe(PO3)2
C2/c
137.51
Genkina et al.54
wNonbridging oxygen bond length and bridging oxygen bond length. zNo exact P–O–P angles reported. yEstimated based on average of reported smallest (1371) and largest
(1491) P–O–P angles.
Figure 2 shows the Raman spectra collected from iron pyro-
phosphate crystalline compounds (O/P5 3.5). The most intense
peaks in each spectrum, between 1000 and 1100 cmꢀ1, corre-
spond to the symmetric PO3 vibrations of inequivalent non-
bridging oxygens associated with the Q1 P2O74ꢀ anions.23 The
less-intense peaks between 1000 and 1200 cmꢀ1 are assigned to
the asymmetric PO3 modes associated with Q1-tetrahedra. The
centrosymmetric Fe2P2O7 crystals do not have as many Raman
active PO3-stretching modes due to the symmetry of
the compound.23,38 Compared with Fe2P2O7 and Fe7(P2O7)4,
the spectra from the crystalline compounds Fe3(P2O7)2 and
Fe4(P2O7)3 have many more peaks assigned to asymmetric
PO3 modes (Fig. 2), consistent with the lower symmetry of these
latter pyrophosphate units (Table III).
for the P5 glass, which has a relatively low O/P ratio (3.16, Table
IV) and so should have a greater fraction of Q2-tetrahedra.
Likewise, the shoulders near B1000 cmꢀ1 could be assigned to
other PO3 modes associated with the Q1-tetrahedra, or to PO4
modes associated with Q0 units in the glasses. The bands be-
tween 700 and 800 cmꢀ1 are related to the symmetric P–O–P-
stretching modes associated with linkages to Q1-tetrahedra. It is
interesting that the peak for the P–O–P symmetric-stretching
mode has a greater relative intensity for the ferrous P2 glass than
for the other glasses. In addition, the peak position varies from
720 cmꢀ1 for the P5 glass to 765 cmꢀ1 for the P2 glass.
Figure 4 shows the Raman spectra of the crystalline ferric and
ferrous metaphosphate (O/P5 3) compounds and several
metaphosphate glasses. The Fe(PO3)3 and Fe(PO3)2 crystalline
compounds have bands at similar frequencies; however, their
relative intensities are quite different. The intense peak near
B1196 cmꢀ1 in the spectrum from the crystalline ferric com-
pound (Fe(PO3)3) corresponds to the PO2-stretching modes
associated with Q2-tetrahedra,15,16,24 and the intense peaks at
1160 and 1205 cmꢀ1 in the spectrum from the crystalline ferrous
compound (Fe(PO3)2) also correspond to these symmetric PO2-
stretching modes associated with inequivalent P–Onb bonds.25
The less intense bands between 1000 and 1250 cmꢀ1 are likely
related to the asymmetric PO2-stretching modes.25,26 The peak
for the symmetric-stretching mode of P–O–P bonds that link
the Q2-tetrahedra in these crystalline compounds is present at
661 cmꢀ1 for Fe(PO3)3 and at 680 cmꢀ1 for Fe(PO3)2; the relative
intensity of the latter peak is significantly greater than the former.
For the glasses in Fig. 4, the major bands between 1150 and
1200 cmꢀ1 are assigned to the PO2-stretching modes associated
with Q2-tetrahedra. The spectra from glasses M1, M2, and M3
each have a broad peak centered near 1060–1070 cmꢀ1. These
peaks are most likely due to PO3-stretching modes associated
with Q1-tetrahedra that terminate phosphate chains. The low-
intensity peak centered near 1300 cmꢀ1 in each spectrum of the
metaphosphate glasses is assigned to asymmetric PO2-stretching
modes associated with Q2-tetrahedra. The peak due to the P–O–
P-stretching modes is present in each spectrum, ranging between
The bands between 700 and 800 cmꢀ1 in Fig. 2 are assigned to
the symmetric P–O–P-stretching mode associated with the
bridging oxygen that links two Q1-tetrahedra in a pyrophos-
phate anion. Crystalline Fe2P2O7 has a relatively intense P–O–P
peak near B731 cmꢀ1, which agrees with the spectrum reported
in Baran et al.23 For the other iron pyrophosphate crystalline
compounds, the intensities of the P–O–P peak are relatively
weak. The peak near B935 cmꢀ1 in the spectrum of Fe7(P2O7)4
is assigned to the asymmetric vibration of P–O–P bonds, as is
the weak peak near B1000 cmꢀ1 in the spectrum of Fe4(P2O7)3.
Figure 3 shows the Raman spectra collected from glasses
with O/P ratio B3.5. These glasses, with different Fe/P ratios
(between 0.65 and 0.83), have similar Raman spectra. In general,
the broad peaks in the spectra from the glasses occur at similar
frequencies as those from the crystalline pyrophosphates
(Fig. 2), and similar assignments can be made for the peaks
in the spectra from the glasses. The most intense peak in each
spectrum, at frequencies from B1060 to B1100 cmꢀ1, can be
assigned to the PO3-stretching modes associated with the non-
bridging oxygens on Q1-tetrahedra.23 The higher frequency
shoulders evident in each spectrum could be due to asymmetric
PO3 modes associated with the Q1-tetrahedra, but also could
be due to symmetric PO2-stretching modes associated with Q2-
tetrahedra. This latter assignment seems particularly obvious
Table IV. Glass Compositions: the O/P Atom Ratios were Calculated from the Fe/P Atom Ratios, Determined by EDS, and the
Fraction of Fe21, Determined by Titration
Glass
O1
O2
O3
P1
P2
P3
P4
P5
Fe/P ratio
1.8070.10 1.0470.15 0.9870.12 0.6870.07 0.9270.13 0.6570.11 0.6770.03 0.4570.02
3571 5071 5173 1671 9574 6478 2271 773
4.8970.10 3.7870.18 3.7270.16 3.4770.12 3.4570.16 3.2770.15 3.4370.04 3.1670.04
Fe21/Fetotal (%)
O/P ratio
Glass
P6
P7
P8
M1
M2
M3
M4
Fe/P ratio
0.7570.08
3072
3.5170.11
0.8370.11
2372
3.6670.16
0.7370.09
16717
3.5370.14
0.3470.03
572
3.0370.04
0.3770.07
1774
3.0270.10
0.5170.06
8174
3.0670.07
0.5170.07
9771
3.0270.07
Fe21/Fetotal (%)
O/P ratio

3126
Journal of the American Ceramic Society—Zhang and Brow
Vol. 94, No. 9
Fig. 1. Raman spectra of iron orthophosphate crystals and glasses.
Fig. 3. Raman spectra of glasses melted from pyrophosphates
compounds.
683 cmꢀ1 (M4) and 720 cmꢀ1 (M1). The relative intensity of this
peak varies considerably with composition.
phosphate glasses fall into similar ranges with the corresponding
iron phosphate crystalline compounds, indicating some structural
similarity between the glassy and corresponding crystalline phos-
phate compounds. For example, the similarity in the Raman spec-
tra from the roller-quenched ‘‘orthophosphate glasses’’ to that
collected from a-FePO4 indicates that the structure of this quartz-
like compound, based on tetrahedral (PO4) and (FeO4) units, is a
better model for the glass structure than the orthorhombic form of
FePO4 with its structure based on (PO4) and (FeO6) units.20
It is obvious that the relative contributions of the various
P–O-stretching modes in the range of 900 and 1400 cmꢀ1 change
with glass composition. To quantify these changes, each spectrum
IV. Discussion
(1) Peak Assignments and Phosphate Tetrahedral
Distributions
Table V summarizes the Raman peak assignments for the various
samples. The assignments are consistent with previous Raman
spectroscopic studies of phosphate crystalline compounds and
glasses.11–34,37,38 The peaks below 600 cmꢀ1 are related to network
bending modes,17,21,37 which are not individually assigned in this
paper. The peak positions vary systematically with phosphate
chemistry, with the frequency (wave numbers) of the P–O-stretch-
ing modes increasing in the order orthophosphateopyro-
phosphateometaphosphate. The Raman frequencies for iron
Fig. 2. Raman spectra of crystalline iron pyrophosphate compounds.
Fig. 4. Raman spectra of iron metaphosphate crystals and glasses.

September 2011
Raman Study of Iron Phosphates
3127
Table V. Summary of Raman Frequency Ranges (cmꢀ1
)
Related to Various Phosphate Groups in Iron Phosphate
Crystalline Compounds and Glasses
Compounds
Glasses
Assignment
200–600
660–680
710–760
960–1010
1010–1100
1120–1200
1150–1210
1260–1300
200–600
680–720
720–780
990–1010
1030–1100
B1200
Network bending
P–O–P symmetric stretch (Q2)
P–O–P symmetric stretch (Q1)
PO4 symmetric stretch (Q0)
PO3 symmetric stretch (Q1)
PO3 asymmetric stretch (Q1)
PO2 symmetric stretch (Q2)
PO2 asymmetric stretch (Q2)
1050–1220
1250–1310
was fit by four Gaussian curves centered near B950–990,
B1040–1090,B1120–1200, and B1250–1320 cmꢀ1. The peak
positions and FWHM of these curves were allowed to vary until
a best-fit solution was achieved. An example of one fit, for the
P4 glass, is shown in Fig. 5.
The relative intensities of the four peaks are plotted as a
function of the O/P ratio for each of the iron phosphate glasses
in Fig. 6. The relative intensities of the peaks near 1180 and near
1300 cmꢀ1 both decrease with increasing O/P ratio and the rel-
ative intensity of the peak centered near 1070 cmꢀ1 increases,
reaching a maximum at O/PB3.5, before decreasing with
greater O/P ratio. The relative intensity of the peak near 970
cmꢀ1 remains low for O/Po3.3, but becomes the dominant
peak for glasses with O/P43.5.
Assignments of these four peaks to specific structural units in
the glass are not unambiguous. It is likely, for example, that the
Raman peak near 1180 cm1 for glasses with O/P ratios between
3.0 and 3.5 will have overlapping contributions from symmetric
PO2-stretching modes associated with P–Onb bonds on Q2-tetra-
hedra, and asymmetric PO3-stretching modes associated with
P–Onb bonds on Q1-tetrahedra; see, for example, the spectra
from crystalline Fe4(P2O7)3 and Fe3(P2O7)2 in Fig. 2, with peaks
at 1150–1200 cmꢀ1 associated with the P2O47ꢀ anions. Never-
theless, the trends in peak intensity in Fig. 6 can be interpreted
using well-known phosphate structural chemistry models if the
peaks near 1180 and 1300 cmꢀ1 are assigned to the symmetric
and asymmetric PO2-stretching modes associated with P–Onb
bonds on Q2-tetrahedra, respectively, the peak centered near
1070 cmꢀ1 is assigned to PO3-stretching modes associated with
P–Onb bonds on Q1-tetrahedra, and the peak centered near
970 cmꢀ1 is assigned to PO4-stretching modes associated with
P–Onb bonds on Q0-tetrahedra. The solid lines in Fig. 6 show the
predicted compositional dependences of the distributions of
Qx-tetrahedra, assuming the simplest chemical model in which
Q2-tetrahedra convert to Q1-tetrahedra, which then convert to
Q0-tetrahedra, with increasing O/P ratio.10 In general, the rela-
tive intensities of the peaks centered at 1180, 1070, and 970 cmꢀ1
change in the same qualitative way as the predicted fractions of
Q2, Q1, and Q0-tetrahedra, respectively. (The compositional
Fig. 6. Relative intensities of the four Gaussian peaks used to fit the
Raman spectra in the range 900–1400 cmꢀ1 from each iron phosphate
glass. The peak positions are (&) 950–990 cmꢀ1, (J) 1040–1090 cmꢀ1
,
( & ) 1120–1200 cmꢀ1, (ꢅ) 1250–1320 cmꢀ1. Solid lines are the predicted
Qx distributions assuming a chemically simple model.10 Dashed lines are
guides for the eye.
dependence of the relative intensity of the peak near 1250–1320
cmꢀ1 parallels that of the 1180 cmꢀ1 peak, supporting the
assignment of the former to asymmetric-vibrational modes on
nonbridging oxygens on Q2-tetrahedra.) Similar spectral trends
have been reported from Raman studies of other polyphosphate
glasses, including Li-phosphates,27 Zn-phosphates,13,28 Ca-, and
Mg-phosphates.29
The Raman spectra of the iron phosphate glasses are more
complex than can be explained by the ‘‘chemically simple’’
structural model, shown in Fig. 6. For example, the spectra col-
lected from glasses like P1, P2, and P3 have peaks or shoulders
that indicate the concomitant presence of Q0-, Q1-, and Q2-
tetrahedra for pyrophosphate compositions that should possess
mostly Q1-tetrahedra. Similar features in the Raman spectra
from Zn-polyphosphate13 and Pb-polyphosphate30 glasses were
interpreted using the Van Wazer site distribution model31 for the
phosphate melts that rely on the following disproportionation
reaction:
2Q12Q2 þ Q0
(1)
It has been shown that increasing the field strength of metal
cations in a polyphosphate melt shifts the site distribution reac-
tion to the right, increasing the structural complexity of the re-
sulting glasses.32,39 Obtaining quantitative measurements of the
site distributions from the Raman spectra is difficult, not least
because of overlapping peaks due to the different symmetric-
and asymmetric-stretching modes associated with the different
tetrahedra. However, the comparison of the relative peak inten-
sities with the predicted Qx-distributions in Fig. 6 is qualitatively
consistent with the site disproportionation model for composi-
tions around the pyrophosphate stoichiometry. It seems clear,
then, that to understand the effects of composition on the prop-
erties of iron phosphate glasses, the role of a broad distribution
of different phosphate anions must be considered.
This interpretation of the Raman spectra of iron phosphate
glasses with compositions around the pyrophosphate stoic-
hiometry is consistent with what has been proposed by others
for similar glasses.9,17,18 That is, Q1-tetrahedra dominate the
phosphate anions that constitute the structures of these glasses,
with Q0-and Q2-tetrahedra also present. The relative concentra-
tions of the latter tetrahedra depend on the O/P ratio of the glass
and may depend on thermal history if reaction (1) is treated as
an equilibrium reaction.31,32
Fig. 5. Decomposition of the Raman spectrum (blue line) for the P4
glass into four Gaussian peaks, the sum of which is indicated by the
dotted line.

3128
Journal of the American Ceramic Society—Zhang and Brow
Vol. 94, No. 9
bond associated with the relevant vibration,13,33 but those stud-
ies involved the effects of cation interactions on the P–Onb
-
stretching modes, not the P–O–P-stretching modes, where the
influence of neighboring ferrous or ferric ions is expected to be
less. It is unclear why the presence of ferrous ions is associated
with an increase in the relative intensity of this peak, although it
is clearly evident for both pyrophosphate and metaphosphate
glasses and crystals. A low intensity (POP)sym peak has been
noted in the Raman spectrum of SrFe2(P2O7)2, but no explana-
tion for this was offered.34
The frequency of the (POP)sym peak is lower for crystalline
iron metaphosphates (e.g., 680 cmꢀ1 for Fe(PO3)2, Fig. 4) than
for the crystalline pyrophosphates (e.g., 731 cmꢀ1 for Fe2P2O7,
Fig. 2). The inset to Fig. 7 shows that for the iron phosphate
glasses, there is a systematic increase in the frequency of this
peak as the O/P ratio increases. These peaks are too broad (and
in some cases, the intensities are too low) to distinguish separate
peaks due to Q1-and Q2-linkages, as can be seen in the Raman
spectra of Zn-13 and Pb-polyphosphate glasses.30 Nevertheless,
this systematic change in the frequency of the (POP)sym peak can
be related to systematic changes in the nature of the P–O–P
linkages, as discussed below.
Fig. 7. Intensity ratios for the P–O–P (600–800 cmꢀ1) and P–Onb (peak
between B1000 and 1200 cmꢀ1) symmetric stretching Raman peaks for
iron phosphate crystalline compounds (Ã) and glasses (ꢅ) as a function
of the fraction of ferrous ions. The inset shows the frequency of the
(POP)sym peak as a function of composition. The dashed line is a guide
for the eye.
(2) P–O Bond Distances
The systematic changes in the Raman peak positions can pro-
vide additional information about the structures of the iron
phosphate glasses and crystalline compounds. Rouse et al.11 in-
dicated that the Raman frequency of the PO2 symmetric-stretch-
ing mode for a series of metaphosphate glasses depends on the
bond force constant between the modifying metal cation and the
nonbridging oxygens (greater force constant, higher Raman
peak frequency), and on the size of the modifying metal cation,
which affects the O–P–O intratetrahedral bond angles (larger
cations increase the bond angle and decrease the Raman peak
frequency). A similar effect can be seen in Fig. 4, where the
PO2 peak positions for the two metaphosphate glasses dominated
by greater field strength ferric ions (M1, 1184 cmꢀ1 and M2, 1205
cmꢀ1) are at greater frequencies than those of the two ‘‘ferrous’’
metaphosphate glasses (M3, 1173 cmꢀ1 and M4, 1170 cmꢀ1).
The Raman spectra from more complex iron phosphate
glasses, for example, those containing other oxides, can be in-
terpreted in a similar manner, if the effects of the field strengths
of other cations are considered. For example, lower-field
strength cations generally shift the Raman peak frequencies
for P–Onb-stretching modes to lower wave numbers,11,12 and
the addition of other oxides may change local symmetries
enough to change the relative intensities of symmetric and asym-
metric peaks from glasses with the same nominal O/P ratio. The
cation effect can be seen in the work of Bingham et al.,18 who
discuss the Raman spectra from K2O- and BaO-containing iron
phosphate glasses. In that study, peaks centered near 470 and
630 cmꢀ1 develop and increase in relative intensity as more
modifying oxides are added to an iron pyrophosphate base
glass, increasing the O/P ratio. There are no comparable peaks
in the ‘‘modifier free’’ compositions in the present study, indi-
cating that the new peaks in the spectra reported by Bingham
and colleagues, and seen in other modified iron phosphate
glasses40 with relatively high O/P ratios (43.5), are due to vib-
rational modes associated with orthophosphate anions that are
charge balanced by the modifying cations. Similar assignments
have been made for the Raman spectra of crystalline alkali iron
orthophosphates.41
Popovic
´
et al.16 has related similar peak shifts to changes in
P–O bond lengths, with shorter bonds corresponding to greater
Raman frequencies. By comparing the Raman spectra from
more than 20 inorganic crystalline phosphates, including ortho-,
pyro-, and metaphosphates, Popovic developed an empirical
´
relationship that correlates the position of the P–O-stretching
mode (n in cmꢀ1) with the P–O bond lengths (R in A)16
n ¼ 6:3 ꢃ 103 ꢀ ð3:43 ꢃ 103ÞR
(2)
Equation (2) produces bond length predictions from Raman
The relative intensity of the P–O–P symmetric-stretching
peak between about 600 and 800 cmꢀ1 appears to depend on
the average oxidation state of the iron ions in these iron phos-
phate glasses and crystalline compounds. For example, compare
the intensity of this peak from the P2 glass (95% Fe21) to that
from the P1 glass (16% Fe21) in Fig. 3. Both glasses have similar
O/P ratios, but the former glass has a much more intense peak at
765 cmꢀ1. Similarly, the relative intensity of the P–O–P sym-
frequencies with uncertainties of 70.01 A for alkali and alkaline
earth phosphates, and Popovic indicated that similar predictions
´
could be made for amorphous materials.
Equation (2) is plotted in Fig. 8, along with the corresponding
average P–O bond lengths (Table III) and the average Raman
peak positions for the iron phosphate crystalline compounds
analyzed in this study. There is good agreement between the
measured and predicted dependences of the Raman peak posi-
tions on the average P–O bond lengths for the P–O-stretching
modes associated with nonbridging oxygens on the Q0-, Q1-, and
Q2-tetrahedra. However, the Raman frequencies for the P–O–P-
stretching modes generally fall below those predicted from the
reported average P–O–P bond lengths of the crystalline pyro-
and metaphosphate compounds. This discrepancy may be due
to the inability to detect and assign the peaks due to asymmetric
P–O–P-stretching modes, which should fall in the range of 900–
1000 cmꢀ1 and so should increase the average peak position to
metric-stretching peak at 683 cmꢀ1 for the M4 glass (97% Fe21
)
is much greater than that for the M1 glass (5% Fe21) (Fig. 4).
Figure 4 also shows that the relative intensity of the P–O–P
symmetric-stretching peak from crystalline ferrous metaphos-
phate, Fe(PO3)2, is significantly greater than for crystalline ferric
metaphosphate, Fe(PO3)3.
Figure 7 plots the ratio of the intensity of the Raman peak
due to the P–O–P symmetric stretch, relative to the most intense
peak due to the P–Onb-stretching modes, as a function of the
Fe21/Fetot ratio; the P–O–P peak intensity increases with an in-
creasing fraction of ferrous ions, particularly when Fe21/Fetot
exceeds B0.6. Changes in Raman peak intensities in phosphate
glasses have been related to the degree of covalency in the P–O
the range predicted by Popovic.
´
Equation (2) was used to predict the average P–Onb and P–
Obr bond distances for the iron phosphate glasses, using the
most intense peak in the range from 1000 to 1300 cmꢀ1 for the

September 2011
Raman Study of Iron Phosphates
3129
that are similar to crystals with similar stoichiometries, although
the glass structures are complicated by the presence of broader
distributions of phosphate anions, produced by disproportion-
ation of chain-terminating (Q1) units to form isolated (Q0) and
chain-linking (Q2) tetrahedra. Systematic changes in Raman
peak positions with glass compositions can be related to changes
in the numbers of bridging and nonbridging oxygens, which lead
to changes in the average P–O bond distances.
Acknowledgments
The authors are very grateful to Prof. Mark E. Schlesinger (Missouri S&T) for
his assistance with the preparation of the crystalline iron phosphate compounds
and Mr. Jong Wook Lim (Missouri S&T) for his help with the collection of the
Raman spectra.
References
1J. E. Miller, M. M. Gonzales, L. Evans, A. G. Sault, C. Zhang, R. Rao, G.
Whitwell, A. Maiti, and D. King-Smith, ‘‘Oxidative Dehydrogenation of Ethane
over Iron Phosphate Catalysts,’’ Appl. Catal. A: Gen., 231 [1] 281–92 (2002).
2M. M. Gadgil and S. K. Kulshreshtha, ‘‘Study of FePO4 Catalyst,’’ J. Solid
State Chem., 111, 357–64 (1994).
Fig. 8. The average positions of the Raman peaks for the P–Onb (J)
and P–Obr (Ã) stretching modes, compared with the average respective
bond distances, for the iron phosphate crystalline compounds studied in
this work. The line is the empirical correlation (Eq. (2)) proposed by
3A. K. Padhi, K. S. Nanjundaswamy, and J. B. Goodenough, ‘‘Phospho-
Olivines as Positive-Electrode Materials for Rechargeable Lithium Batteries,’’
J. Electrochem. Soc., 144 [4] 1188–94 (1997).
Popovic
´
et al.16
former, and the peak due to the P–O–P-symmetric stretch in the
600–800 cmꢀ1 range for the latter. These bond distances are re-
ported as a function of O/P ratio in Fig. 9. In general, it appears
that the average P–Onb bonds become longer and the average P–
Obr bonds become shorter as the O/P ratio increases. Also plot-
ted are the overall average P–O bond lengths, calculated from a
weighted average of P–Onb and P–Obr distances for each respec-
tive glass. The overall average P–O bond distance decreases
slightly, from 1.57 to 1.54 A, with increasing O/P ratio, as non-
bridging oxygens replace bridging oxygens. The open symbols in
Fig. 9 report the average P–O bond distances from X-ray and
neutron diffraction studies of ferrous metaphosphate42 and fer-
ric polyphosphate glasses.43,44 Overall, there is reasonable agree-
ment between the P–O bond distances predicted from the
Raman spectra and those reported from the diffraction studies,
although the former are consistently B0.02 A longer than the
latter. This difference is likely associated with the assumptions
4A. Yamada, S. C. Chung, and K. Hinokuma, ‘‘Optimized LiFePO4 for Lithium
Battery Cathodes,’’ J. Electrochem. Soc., 148, A224–9 (2001).
5Y.-S. Hong, Y. J. Park, K. S. Ryu, and S. H. Chang, ‘‘Crystalline Fe3PO7 as an
Electrode Material for Lithium Secondary Batteries,’’ Solid State Ion., 156 [1-2]
27–33 (2003).
6M. G. Mesko and D. E. Day, ‘‘Immobilization of Spent Nuclear Fuel in Iron
Phosphate Glass,’’ J. Nucl. Mater., 273, 27–36 (1999).
7X. Yu, D. E. Day, G. J. Long, and R. K. Brow, ‘‘Properties and Structure of
Sodium–Iron Phosphate Glasses,’’ J. Non-Cryst. Solids, 215, 21–31 (1997).
8G. K. Marasinghe, M. Karabulut, C. S. Ray, D. E. Day, M. G. Shumsky, W.
B. Yelon, C. E. Booth, P. G. Allen, and D. K. Shuh, ‘‘Structural Features of Iron
Phosphate Glasses,’’ J. Non-Cryst. Solids., 222, 144–52 (1997).
9X. 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).
10R. K. Brow, ‘‘Review: The Structure of Simple Phosphate Glasses,’’ J. Non-
Cryst. Solids, 263/264, 1–28 (2000).
11G. B. Rouse, P. J. Miller, and W. M. Risen, ‘‘Mixed Alkali Glass Spectra and
Structure,’’ J. Non-Cryst. Solids, 28, 193–207 (1978).
12B. N. Nelson and G. J. Exarhos, ‘‘Vibrational Spectroscopy of Cation-Site
Interactions in Phosphate Glasses,’’ J. Chem. Phys., 71 [7] 2739–47 (1979).
13R. K. Brow, D. R. Tallant, S. T. Myers, and C. C. Phifer, ‘‘The Short Range
Structure of Zinc Phosphate Glass,’’ J. Non-Cryst. Solids, 191, 45–55 (1995).
14J. J. Hudgens, R. K. Brow, D. R. Tallant, and S. W. Martin, ‘‘Raman Spec-
troscopy Study of the Structure of Lithium and Sodium Ultraphosphate Glasses,’’
J. Non. Cryst. Solids, 223 [1,2] 21–31 (1998).
made in the use of the Popovic relationship.
´
V. Summary
Raman spectra collected from crystalline ferrous, ferric, and
ferrous-ferric phosphate compounds and glasses provide infor-
mation about the phosphate anions that constitute the struc-
tures of these materials. In general, the glasses have structures
15A. Rulmont, R. Cahay, M. Liegeois-Duyckaerts, and P. Tarte, ‘‘Vibrational
Spectroscopy of Phosphates: Some General Correlations between Structure and
Spectra,’’ Eur. J. Solid State Inorg. Chem., 28, 207–19 (1991).
16L. Popovic
´
, D. de Waal, and J. C. A. Boeyens, ‘‘Correlation between Raman
Wavenumbers and P–O Bond Lengths in Crystalline Inorganic Phosphates,’’
J. Raman Spectrosc., 36, 2–11 (2005).
17A. Mogus-Milankovic, B. Pivac, K. Furic, and D. E. Day, ‘‘Structural Study
of Iron Phosphate Glasses,’’ Phys. Chem. Glasses, 38 [2] 74–8 (1997).
18P. A. Bingham, R. J. Hand, O. M. Hannant, S. D. Forder, and S. H. Kilcoyne,
‘‘Effects of Modifier Additions on the Thermal Properties, Chemical Durability,
Oxidation State and Structure of Iron Phosphate Glasses,’’ J. Non-Cryst. Solids,
355 [28–30] 1526–38 (2009).
19L. 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).
20C. 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).
21M. P. Pasternak, G. Kh.Rozenberg, A. P. Milner, M. Amanowicz, T. Zhou,
U. Schwarz, K. Syassen, R. Dean Miller, M. Hanfland, and K. Brister, ‘‘Pressure-
induced Concurrent Transformation to an Amorphous and Crystalline Phase in
Berlinite-Type FePO4,’’ Phys. Rev. Lett., 79 [22] 4409–12 (1997).
22W. H. Baur, ‘‘The Geometry of Polyhedral Distortions. Predictive Relation-
ships for the Phosphate Group,’’ Acta Cryst. B, 30, 1195–215 (1974).
23E. J. Baran, I. L. Botto, and A. G. Nord, ‘‘The Vibrational Spectrum and the
Conformation of the P2O47ꢀ Anion in Fe2P2O7,’’ J. Mol. Struct., 143, 151–4 (1986).
24S. H. Morgan, R. H. Magruder, and E. Silberman, ‘‘Raman Spectra of Rare-
Earth Phosphate Glasses,’’ J. Am. Ceram. Soc., 70, C378–80 (1987).
25K. Viswanathan, V. U. Nayar, and G. Aruldhas, ‘‘Infrared and Raman Spec-
tra of Three Tetrametaphosphates M2P4O12 (M 5 Fe, Ni, Zn),’’ J. Chem. Sci., 95
[5–6] 463–9 (1985).
Fig. 9. P–O bond distances (closed symbols) for the iron phosphate
glasses calculated from Raman peak positions using Eq. (2). The open
symbols are bond distances reported in the literature.42–44 Dashed lines
are guides for the eye.
26W. Zhou, W. He, X. Zhang, J. Liu, Y. Du, S. Yan, X. Tian, X. Sun, X. Han,
and Y. Yue, ‘‘Simple and Rapid Synthesis of Fe(PO3)3 by Microwave Sintering,’’
J. Chem. Eng. Data, 54, 2073–6 (2009).

3130
Journal of the American Ceramic Society—Zhang and Brow
Vol. 94, No. 9
27M. Tatsumisago, Y. Kowada, and T. Minami, ‘‘Raman Spectra of Rapidly
42U. Hoppe, M. Karabulut, E. Metwalli, R. K. Brow, and P. Jo
vari, ‘‘The Fe–O
´ ´
Quenched Glasses and Melts Containing Large Amounts of Li2O,’’ J. Non-Cryst.
Solids, 150, 207–11 (1992).
Coordination in Iron Phosphate Glasses by X-Ray Diffraction with High Energy
Photons,’’ J. Phys.: Condens. Matter, 15, 6143–53 (2003).
28H. Takebe, Y. Baba, and M. Kuwabara, ‘‘Dissolution Behavior of ZnO–P2O5
Glasses in Water,’’ J. Non-Cryst. Solids, 352, 3088–94 (2006).
29M. A. Karakassides, A. Saranti, and I. Koutselas, ‘‘Preparation and Structural
Study of Binary Phosphate Glasses with High Calcium and/or Magnesium Con-
tent,’’ J. Non-Cryst. Solids, 347 [1–3] 69–79 (2004).
43M. Karabulut, G. K. Marasinghe, C. S. Ray, G. D. Waddill, and D. E. Day,
‘‘A High Energy X-ray and Neutron Scattering Study of Iron Phosphate Glasses
Containing Uranium,’’ J. Appl. Phys., 87, 2185–93 (2000).
44A. C. Wright, R. N. Sinclair, J. L. Shaw, R. Haworth, G. K. Marasinghe, and
D. E. Day, ‘‘A Neutron Diffraction Study of the Structure of Iron Phosphate
Glasses,’’ Phys. Chem. Glasses Eur. J. Glass Sci. Technol. B, 49 [1] 1–7 (2008).
45A. Goiffon, J. C. Dumas, and E. Philippot, ‘‘Alpha-Quartz Type Phases:
Structure of FePO4 and [57]Fe Mossbauer Spectroscopy,’’ Rev. Chim. Miner., 23,
99–110 (1986).
30G. LeSaout, P. Simon, F. Fayon, A. Blin, and Y. Vaills, ‘‘Raman and Infrared
Study of (PbO)x(P2O5)(1ꢀx) Glasses,’’ J. Raman Spectrosc., 33, 740–6 (2002).
31J. R. Van Wazer, Phosphorus and its Compounds. Interscience, New York, NY,
1958.
32T. R. Meadowcraft and F. D. Richardson, ‘‘Structural and Thermodynamic
Aspects of Phosphate Glasses,’’ Trans. Faraday Soc., 61, 54–70 (1965).
33J. Koo, B.-S. Bae, and H.-K. Na, ‘‘Raman Spectroscopy of Copper Phosphate
Glasses,’’ J. Non-Cryst. Solids, 212, 173–9 (1997).
46J. K. Warner, A. K. Cheetham, A. G. Nord, R. B. Von Dreele, and M. Yeth-
iraj, ‘‘Magnetic Structure of Iron(II) Phosphate, Sarcopside, Fe3(PO4)2,’’ J. Mater.
Chem., 2, 191–6 (1992).
47J. K. Warner, A. K. Cheetham, and D. E. Cox, ‘‘Determination of Cation
Distribution in NiFe2(PO4)2 Using Resonant X-ray and Neutron Powder Diffrac-
tion,’’ J. Appl. Cryst., 28, 494–502 (1995).
34E. J. Baran, R. C. Mercader, A. Massaferro, and E. Kremer, ‘‘Vibrational and
57Fe-Mo
¨
ssbauer Spectra of Some Mixed Cation Diphosphates of the Type
MIIFe2III(P2O7)2,’’ Acta Part A, 60, 1001–5 (2004).
48Y. A. Gorbunov, B. A. Maksimov, Y. K. Kabalov, A. N. Ivashchenko, O. K.
Mel’nikov, and N. V. BeloV, ‘‘Crystal Structure of Fe231Fe431 [PO4]6,’’ Dokl. Akad.
Nauk SSR, 254, 873–7 (1980).
35L. Zhang, M. E. Schlesinger, and R. K. Brow, ‘‘Phase Equilibria in the Fe2O3–
P2O5 System,’’ J. Am. Ceram. Soc., doi: 10.1111/j.1551-2916.2010.04287.x (2011).
36S. I. Grishin, J. M. Bigham, and O. H. Tuovinen, ‘‘Characterization of Jaro-
site Formed upon Bacterial Oxidation of Ferrous Sulfate in a Packed Bed Reac-
tor,’’ Appl. Environ. Microbiol., 54, 3101–6 (1988).
49L. K. Elbouaanani, B. Malaman, R. Ge
´
ture Refinement and Magnetic Properties of Fe4(P2O7)3 Studied by Neutron
rardin, and M. Ijjaali, ‘‘Crystal Struc-
Diffraction and Mossbauer Techniques,’’ J. Solid State Chem., 163, 412–20 (2002).
50J. T. Hoggins, J. S. Swinnea, and H. Steinfink, ‘‘Crystal Structure of
Fe2P2O7,’’ J. Solid State Chem., 47, 278–83 (1983).
¨
37R. L. Frost, T. Kloprogge, M. L. Weier, W. N. Wayde, Z. Ding, and G. H.
Edwards, ‘‘Raman Spectroscopy of Selected Arsenates—Implications for Soil Re-
mediation,’’ Spectrochim. Acta, Part A, 59, 2241–6 (2003).
51M. Ijjaali, G. Venturini, R. Gerardin, B. Malaman, and C. Gleitzer, ‘‘Synthe-
sis Structure and Physical Properties of a Mixed-Valence Iron Diphosphate
Fe3(P2O7)2: First Example of Trigonal Prismatic Fe21 with O2-Ligands,’’ Eur. J.
Solid State Inorg. Chem., 28, 983–98 (1991).
38K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination
Compounds. John Wiley & Sons Inc., New York, NY, (1997).
39B. C. Sales and L. A. Boatner, ‘‘Physical and Chemical Characteristics of
Lead-Iron Phosphate Nuclear Waste Glasses,’’ J. Non-Cryst. Solids, 79, 83–116
(1986).
52B. Malaman, M. Ijjali, R. Gerardin, G. Venturini, and C. Gleitzer,
‘‘Fe7(P2O7)4, a Mixed Valence Iron Diphosphate the Missing Link Between
Fe2P2O7 and Fe3(P2O7)2,’’ Eur. J. Solid State Inorg. Chem., 29, 1269–84 (1992).
53J. M. Rojo, J. L. Mesa, L. Lezama, and T. Rojo, ‘‘Magnetic Properties of the
Fe(PO3)3 Metaphosphate,’’ J. Solid State Chem., 145, 629–33 (1999).
54E. A. Genkina, B. A. Maksimov, and O. K. Melnikov, ‘‘Crystal Structure of
Iron Tetrametaphosphate Fe2P4O12 and its Comparison with the Structure of Other
40A. Mogug-Milankovic
, A. Santic, S. T. Reis, K. Furic, and D. E. Day, ‘‘Stud-
´ ´ ´
ies of Lead–iron Phosphate Glasses by Raman, Mossbauer and Impedance Spec-
¨
troscopy,’’ J. Non-Cryst. Solids, 351, 3246–58 (2005).
41G. Butt, N. Sammes, G. Tompsett, A. Smirnova, and O. Yamamoto, ‘‘Raman
Spectroscopy of Superionic Ti-Doped Li3Fe2(PO4)3 and LiNiPO4 Structures,’’
J. Power Sources, 134, 72–9 (2004).
M221 PO4(M21 5 Ni, Mg, Cu, Co),’’ Sov. Phys. Crystallogr., 30, 885–9 (1985).
&