Crystal growth, single-crystal structure refinement and unusual ligand-field splittings of lazulite-type oxidephosphates MTi2O 2(PO4)I (M = FeII, CoII, Ni II)

Article abstract of DOI:10.1002/zaac.200700293

Single crystals of oxidephosphates MTi₂O₂(PO ₄)₂ [M: Fe (dark red), Co (pinkish red), Ni (green)] with edge-lengths up to 0.4 mm were grown by chemical vapour transport. FeTi ₂O₂(PO₄)₂ and CoTi₂O ₂(PO₄)₂ are isotypic to NiTi₂O ₂(PO₄)₂. The crystal structure of the latter was previously solved from powder data [FeTi₂O₂(PO ₄)₂ (data for CoTi₂O₂(PO ₄)₂ and NiTi₂O₂(PO₄) ₂ in brackets): monoclinic, P2₁/c, Z = 2, a = 7.394(3) (7.381(6), 7.388(4)) A, b = 7.396(2) (7.371(5), 7.334(10))A, c = 7.401(3) (7.366(6), 7.340(3)) A, β = 120.20(3) (120.26(6), 120.12(4))°, R₁ = 0.0393 (0.0309, 0.0539) wR₂ = 0.1154 (0.0740, 0.1389), 2160 (1059, 1564) independent reflections, 75 (76, 77) variables]. The single-crystal study allowed improved refinement using anisotropic displacement parameters, yielded lower standard deviations for the structural parameters and revealed a small amount of cation disordering. Twinning and cation disordering within the structures are rationalized by a detailed crystallographic classification of the MTi₂O ₂(PO₄)₂ structure type in terms of group-subgroup relations. The structure is characterized by a three-dimensional network of [PO₄] tetrahedra and [M^IITi₂O ₁₂] groups formed by face-sharing of [M^IIO₆] and [TiO₆] octahedra. Electronic absorption spectra of MTi ₂O₂(PO₄)₂ in the UV/VIS/NIR region show rather large ligandfield splittings for the strongly trigonally distorted chromophors [M^IIO₆] (M = Fe, Co, Ni) with interelectronic repulsion parameters beeing slightly smaller than in other phosphates. Interpretation of the spectra within the framework of the angular overlap model reveals a significant second-sphere ligand field effect of Ti^IV ions on the electronic levels of the Ni^II and Co^II.

Full text of DOI:10.1002/zaac.200700293

DOI: 10.1002/zaac.200700293
Crystal Growth, Single-Crystal Structure Refinement and Unusual Ligand-Field
Splittings of Lazulite-Type Oxidephosphates MTi2O2(PO4)2 (M ؍ FeII, CoII, NiII)
M. Schöneborn [1] and R. Glaum*
Bonn, Institut für Anorganische Chemie der Universität
Received June 12th, 2007.
Dedicated to Professor Rudolf Hoppe on the Occasion of his 85th Birthday
Abstract. Single crystals of oxidephosphates MTi2O2(PO4)2 [M: Fe
(dark red), Co (pinkish red), Ni (green)] with edge-lengths up to
0.4 mm were grown by chemical vapour transport. FeTi2O2(PO4)2
and CoTi2O2(PO4)2 are isotypic to NiTi2O2(PO4)2. The crystal
structure of the latter was previously solved from powder data
[FeTi2O2(PO4)2 (data for CoTi2O2(PO4)2 and NiTi2O2(PO4)2 in
brackets): monoclinic, P21/c, Z ϭ 2, a ϭ 7.394(3) (7.381(6),
rationalized by a detailed crystallographic classification of the
MTi2O2(PO4)2 structure type in terms of group-subgroup relations.
The structure is characterized by a three-dimensional network of
[PO4] tetrahedra and [MIITi2O12] groups formed by face-sharing
of [MIIO6] and [TiO6] octahedra. Electronic absorption spectra of
MTi2O2(PO4)2 in the UV/VIS/NIR region show rather large ligand-
field splittings for the strongly trigonally distorted chromophors
[MIIO6] (M ϭ Fe, Co, Ni) with interelectronic repulsion parameters
beeing slightly smaller than in other phosphates. Interpretation of
the spectra within the framework of the angular overlap model re-
veals a significant second-sphere ligand field effect of TiIV ions on
the electronic levels of the NiII and CoII.
˚
˚
7.388(4)) A, b ϭ 7.396(2) (7.371(5), 7.334(10)) A, c ϭ 7.401(3)
˚
(7.366(6), 7.340(3)) A, β ϭ 120.20(3) (120.26(6), 120.12(4))°, R1 ϭ
0.0393 (0.0309, 0.0539) wR2 ϭ 0.1154 (0.0740, 0.1389), 2160 (1059,
1564) independent reflections, 75 (76, 77) variables]. The single-
crystal study allowed improved refinement using anisotropic dis-
placement parameters, yielded lower standard deviations for the
structural parameters and revealed a small amount of cation disor-
dering. Twinning and cation disordering within the structures are
Keywords: Chemical vapour transport; Crystal growth; Electronic
absorption spectra; Lazulite; Ligand-field effects
Introduction
aristotype [9] of the Lipscombite/Lazulite structure family,
tetragonal β-Fe2O(PO4) [10, 11, 12], and several ternary and
polynary titanium phosphates. Occurence of disorder and
twinning within the structures of this family are almost
inevitable and can be linked to these symmetry relations.
The particular unit cell dimensions found for oxide
phosphates MTi2O2(PO4)2 lead to additional pseudo-
merohedral twinning which complicates single-crystal struc-
The crystal structure of NiTi2O2(PO4)2 („Ni0.5TiPO5” in
[2]) has been solved and refined some years ago from X-ray
powder diffraction data, in addition the electronic absorp-
tion spectrum (powder reflectance measurement) has been
reported [2]. Only recently, structural data for FeTi2-
O2(PO4)2 from powder data [3] and for two modifications
of CuTi2O2(PO4)2 [4, 5] were published. These oxidephos-
phates belong to the Lipscombite/Lazulite structure family
[6, 7]. Accomodation of various cations in different oxi-
dation states and the possible substitution of O2Ϫ by OHϪ
and H2O illustrate the wide crystal chemical stability range
of this structure type. Table 1 reviews phosphates of the
Lipscombite/Lazulite family. The crystal chemical varia-
bility of this structure family demands for a structural sys-
tematization. Symmetry considerations on the basis of
group-subgroup relations [8] provide a valuable tool for this
purpose. Figure 1 shows the symmetry relation between the
Table 1 Phosphates of the Lipscombite/Lazulite structure family.
ᮀ means octahedral void.
phosphate
Z
space group reference
β-FeII/III2O(PO4); aristotype
β-VII/III2O(PO4)
4
4
4
2
I41/amd
I41/amd
I41/amd
P21/c
[10, 11]
[53]
[11]
[2], this paper
[54, 1]
[54, 55, 56]
[54]
[57]
[57]
[58, 59, 60]
[58]
[58]
NiIICrIIIO(PO4)
ᮀMIITiIV2O2(PO4)2 (M: Fe, Co, Ni)
ᮀ2TiIIITiIV3O3(PO4)3
ᮀ3TiIV5O4(PO4)4
24 F2dd
4
1
4
P212121
F2dd
P21/n
ᮀ
17CrIII4TiIV27O24(PO4)24
ᮀVIII(H2O)PO4
III
ᮀ
0.385
V
1.23(OH)0.69(H2O)0.31(PO4)·0.33 H2O 4 I41/amd
ᮀFeIIFeIII2(OH)2(PO4)2
ᮀFeIIAlIII2(OH)2(PO4)2
ᮀMgIIAlIII2(OH)2(PO4)2
ᮀCuIIFeIII2(OH)2(PO4)2
ᮀFeIII4(OH)3(PO4)3
2
2
2
2
4
2
2
P21/n
P21/n
P21/n
P21/n
C2/c
I41/amd
P21/c
* Prof. Dr. R. Glaum
[61]
Institut für Anorganische Chemie, Uni Bonn
Gerhard-Domagk-Straße 1
D-53121 Bonn
[62, 63]
[64, 65]
[13, 66]
ᮀ
2ϪxFeIII4ϪxFeII3x(OH)3ϪxO3x(PO4)3
ᮀCo3(OH)2(HPO4)2
e-mail: rglaum@uni-bonn.de
2568
 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2007, 633, 2568Ϫ2578

Lazulite-Type Oxidephosphates MTi2O2(PO4)2 (M ϭ FeII, CoII, NiII)
Fig. 1 Group-subgroup relations within the symmetry tree of the Lipscombite/Lazulite structure family.
TiO2 and MO (M: Co, Ni) (eq. (1)) or “FeO” (1/3 Fe2O3 ϩ 2/3 Fe)
ture refinements even more. Similar problems were encoun-
tered during structure determination and refinement of
Co3(OH)2(HPO4)2 [13].
(eq. (2)). TiP2O7 was prepared as described in [18]. The other start-
ing materials were at least of p. a. grade. Their crystal chemical
composition was checked by Guinier photographs.
In addition to the crystallographic problems growth of
single crystals of polynary phosphates is in general not an
easy task. Even more demanding is crystallization of phos-
phates containing transition metals in rather low oxidation
states like iron(II). In recent years we could show that the
method of chemical vapour transport [14] is well suited for
the crystallization of anhydrous (transition metal) phos-
phates [15, 16]. We therefore applied this technique to crys-
tal growth of oxide phosphates MTi2O2(PO4)2 (M ϭ Fe,
Co, Ni, Cu, Zn).
The green colour of NiTi2O2(PO4)2 and the unusual UV/
vis spectrum reported in literature [2] are in marked con-
trast to yellow and orange colours observed for many other
anhydrous nickel(II) phosphates [17]. Thus our interest in
the ligand-field splitting experienced by the divalent cations
in these structures was raised.
TiP2O7 ϩ TiO2 ϩ MO ϭ MTi2O2(PO4)2 (M: Co, Ni)
3 TiP2O7 ϩ 3 TiO2 ϩ Fe2O3 ϩ 2 Fe ϭ 3 FeTi2O2(PO4)2
(1)
(2)
Chemical vapour transport experiments for crystallisation and puri-
fication of the oxidephosphates MTi2O2(PO4)2 (M ϭ Fe, Co, Ni)
were carried out in sealed silica tubes (l ϳ 12 cm, d ϳ 1.5 cm,
V ϳ 21 cm3), which were placed in the temperature gradient of
electrically heated two-zone furnaces. Various transport agents
were checked [chlorine (from in situ thermal decomposition of
PtCl2), mixtures H2/HCl/(N2) (from in situ thermal decomposition
of NH4Cl), HCl/(N2) (from in situ thermal decomposition of stoi-
chiometric amounts of PtCl2 and NH4Cl), mixtures I2/PI3/TiI4
(from in situ reaction of TiP with a surplus of iodine), and mixtures
Cl2/PCl3/TiCl4 (from in situ reaction of TiP with a surplus of chlor-
ine)]. Generally, stoichiometric mixtures of TiP2O7, TiO2 and MO
(M: Co, Ni) were used as starting materials. Experiments with sep-
arately synthesized oxidephosphates MTi2O2(PO4)2 as source mate-
rials did not show any siginificant differences in their transport
behaviour. Conditions and results of representative experiments are
summarized in Tables 2 to 4. Figure 2 shows crystals of the oxide-
phosphates obtained by CVT.
Experimental Section
Synthesis. Microcrystalline powders of MTi2O2(PO4)2 (M ϭ Fe,
Co, Ni) were obtained by isothermal heating in sealed silica tubes
(3 days, 1000 °C) of equimolar amounts of finely ground TiP2O7,
Z. Anorg. Allg. Chem. 2007, 2568Ϫ2578
 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.zaac.wiley-vch.de
2569

M. Schöneborn, R. Glaum
Table 2 Chemical vapour transport experiments for the crystalli-
zation of NiTi2O2(PO4)2. Starting materials in all experiments:
TiP2O7 (235.7 mg, 1.0617 mmol), TiO2 (84.4 mg, 1.0617 mmol)
und NiO (79.3 mg, 1.0617 mmol), TA: transport agent, T2 Ǟ T1:
temperature gradient, yield of crystals at the sink region of the
ampoule: some mg (ϩ), some crystals (ϳ), few crystals (Ϫ).
Table 4 Chemical vapour transport experiments for the crystalli-
zation of FeTi2O2(PO4)2. Starting materials in all experiments:
TiP2O7 (237.34 mg, 1.07 mmol), TiO2 (85.45 mg, 1.07 mmol), Fe
(19.92 mg, 0.36 mmol) and Fe2O3 (56.96 mg, 0.36 mmol), TA:
transport agent, T2 Ǟ T1: temperature gradient, yield of crystals
at the sink region of the ampoule: some mg (ϩ), some crystals (ϳ),
few crystals (Ϫ).
TA
/ mg
T2 Ǟ T1
/ °C
t
products
/ d source
sink
TA
/ mg
T2 Ǟ T1
/ °C
t
/ d
products
source
sink
PtCl2
PtCl2
PtCl2
NH4Cl
30
50
50
5
1000 Ǟ 900
1000 Ǟ 900
1000 Ǟ 900
7
9
7
NiTi2O2(PO4)2 no transport
NiTi2O2(PO4)2 no transport
NiTi2O2(PO4)2 no transport
NH4Cl
30
1000 Ǟ 900
1000 Ǟ 900
7
7
FeTi2O2(PO4)2
FeTi4(PO4)6 (Ϫ)
Fe2P2O7 (ϳ)
PtCl2
NH4Cl
PtCl2
TiP
I2
TiP
50
15
FeTi2O2(PO4)2
FeTi4(PO4)6
FeTi2O2(PO4)2
FeTi4(PO4)6 (Ϫ)
Fe2P2O7 (Ϫ)
NH4Cl 15
1000 Ǟ 900
1000 Ǟ 900
7
NiTi2O2(PO4)2 Ti4P6Si2O25 (ϳ)
NiTi2O2(PO4)2 (Ϫ)
NiTi2O2(PO4)2 no transport
a, b)
100 1000 Ǟ 900 14
2.5
95
05
FeTi2O2(PO4)2
I2
TiP
PtCl2
TiP
PtCl2
TiP
PtCl2
TiP
PtCl2
TiP
PtCl2
TiP
PtCl2
TiP
95
5
7
1000 Ǟ 900
7
FeTi2O2(PO4)2
no transport
100 1000 Ǟ 900
1.5
100 1000 Ǟ 900
2.5
100 1060 Ǟ 960
2.5
100 1060 Ǟ 960
2.5
100 1020 Ǟ 920,
2.5
7
NiTi2O2(PO4)2 NiTi2O2(PO4)2 (Ϫ)
NiTi2O2(PO4)2 NiTi2O2(PO4)2 (Ϫ)
NiTi2O2(PO4)2 NiTi2O2(PO4)2 (ϩ)
NiTi2O2(PO4)2 NiTi2O2(PO4)2 (Ϫ)
7
a) The crystal for the single-crystal data collection was selected from this
batch. b) An excess of TiO2 was added (5 mg).
14
4
By EDX-analysis the chemical composition of deposited crystals
was confirmed (scanning tunneling microscope DMS 940, Zeiss
Inc., equipped with an energy-dispersive spectrometer, EDAX
Inc.). No incorporation of silicon into the phosphates was detected.
a)
1
3
7
NiTi2O2(PO4)2 NiTi2O2(PO4)2
1000 Ǟ 900
100 1000 Ǟ 900
5
NiTi2O2(PO4)2 NiTi2O2(PO4)2 (ϩ)
Ͱ-SiO2
a) The crystal for the single-crystal data collection was selected from this
experiment.
Single crystal electronic spectra of arbitrary faces of MTi2O2(PO4)2
(M ϭ Fe, Co, Ni) crystals were measured at ambient temperature
using a strongly modified CARY 17 microcrystal spectralpho-
tometer (Spectra Services, ANU Canberra, Australia [19, 20]). The
spectrometer allows measurement of polarised spectra of very small
single-crystals with diameters down to 0.1 mm. Details on the spec-
trometer have already been described in literature [19, 20].
Table 3 Chemical vapour transport experiments for the crystalli-
zation of CoTi2O2(PO4)2. Starting materials in all experiments:
TiP2O7 (235.6 mg, 1.0619 mmol), TiO2 (84.4 mg, 1.0617 mmol)
und CoO (79.6 mg, 1.0619 mmol), TA: transport agent, T2 Ǟ T1:
temperature gradient, yield of crystals at the sink region of the
ampoule: some mg (ϩ), some crystals (ϳ), few crystals (Ϫ).
Guinier photographs using the image plate technique [21, 22] were
taken for identification, examination of the powder samples for pu-
rity and for determination of lattice parameters. Details on this
technique were reported in [23]. The lattice parameters were deter-
mined using the program SOS [24].
TA
/ mg
T2 Ǟ T1
/ °C
t
products
X-Ray single-crystal study. Intensity data were measured on a κ-
CCD diffractometer (Enraf-Nonius Inc.) at ambient temperature.
Structure determination and refinement were performed using the
SHELX-97 [25] suite in the WinGX [26] framework. In addition
the program KPLOT [27] was used to search for missing symmetry.
Starting parameters for the single crystal structure refinements
/ d source
sink
PtCl2
PtCl2
55
55
1000 Ǟ 900
7
CoTi2O2(PO4)2 no transport
1060 Ǟ 960 10 CoTi2O2(PO4)2 CoTi2O2(PO4)2 (Ϫ)
CoTi4(PO4)6 (ϩ)
Ͱ-Co2P2O7 (ϩ)
a, b)
PtCl2
PtCl2
105 1040 Ǟ 940
100 1040 Ǟ 940
7
7
7
CoTi2O2(PO4)2 CoTi2O2(PO4)2 (ϩ)
CoTi4(PO4)6 (Ϫ)
CoTi2O2(PO4)2 CoTi4(PO4)6 (ϩ)
Ͱ-Co2P2O7 (ϩ)
CoTi2O2(PO4)2 CoTi2O2(PO4)2 (Ϫ)
CoTi4(PO4)6 (ϩ)
¯
were obtained by Direct Methods assuming space group P1. This
turned out to be necessary, because of twinning of most crystals,
no reliable space-group determination was possible on the basis of
systematic absences and statistics of intensity distribution. Thus,
the metal and phosphorus atoms were revealed. Subsequent ∆-
Fourier syntheses allowed localisation of the oxygen atoms. All
atoms were refined allowing for anisotropic displacement param-
eters. Afterwards the search for higher symmetry was made using
KPLOT. P21/c was found as possible space group in agreement with
the structure model derived from x-ray powder data for NiTi2-
O2(PO4)2 [2]. After transformation of the lattice and the atomic
parameters the twinnig law [1 0 0 0 -1 0 -1 0 -1] was introduced for
CoTi2O2(PO4)2 and NiTi2O2(PO4)2. This twinning law was derived
by group-subgroup considerations [8], relating the structures of the
polynary oxidephosphates to that of the tetragonal aristotype β-
Fe2O(PO4) (Fig. 1). Fig. 4 visualizes the relation between the mon-
oclinic unit cell and an orthorhombic F-lattice. A two-fold axis
PtCl2
NH4Cl
50
5
1000 Ǟ 900
Ͱ-Co2P2O7 (ϳ)
PtCl2
NH4Cl
NH4Cl 15
100 1000 Ǟ 900 14 CoTi2O2(PO4)2 CoTi4(PO4)6 (ϩ)
5
CoTi4(PO4)6
1000 Ǟ 900 14 CoTi2O2(PO4)2 CoTi4(PO4)6 (ϩ)
CoTi4(PO4)6 Ͱ-Co2P2O7 (Ϫ)
CoTi2O2(PO4)2 CoTi4(PO4)6 (ϩ)
CoTi4(PO4)6 Ͱ-Co2P2O7 (ϳ)
CoTi2O2(PO4)2 CoTi4(PO4)6 (ϩ)
CoTi4(PO4)6
Ͱ-Co2P2O7 (ϳ)
I2
TiP
I2
TiP
PtCl2
TiP
95
5
1000 Ǟ 900
7
7
7
100 1000 Ǟ 900
1.5
Ͱ-Co2P2O7 (ϳ)
100 1000 Ǟ 900
CoTi2O2(PO4)2 CoTi2O2(PO4)2 (ϳ)
CoTi4(PO4)6
Ͱ-SiO2
1.5
a) The crystal suitable for the single-crystal data collection was selected from
this batch.
b) An excess of TiO2 was added (5 mg).
2570
www.zaac.wiley-vch.de
 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2007, 2568Ϫ2578

Lazulite-Type Oxidephosphates MTi2O2(PO4)2 (M ϭ FeII, CoII, NiII)
Fig. 3 Electronic single-crystal absorption spectra of oxidephos-
phates MTi2O2(PO4)2 (M: Fe, Co, Ni). The spectra are shifted
along the ordinate axis for clarity. Hairlines are meant as an guide
to the eye. Ticks (a, b, c, d) represent transition energies calculated
within the AOM. The length of the ticks is proportional to the
degree of triplett (for NiII) and quartet (for CoII) character, respec-
tively, of the corresponding excited state.
Fig. 2 Crystals of oxidephosphates MTi2O2(PO4)2 [M: Fe (a), Co
(b), Ni (c)] from chemical vapour transport experiments. Edge-
lengths up to 0.4 mm.
present in the orthorhombic structure is preserved as twin oper-
ation. For all three oxidephosphates remaining electron density at
special postion 2d (site Fe2, at the centre of an octahedral void,
sharing opposite faces with adjacent octahedra [TiO6]; Fig. 6) and
at general position 4e (site Ti2, close to the position of Ti1; Figure
6) indicated small amounts of disorder. Eventually, in the structure
models partial occupancy of both positions 2a and 2d was assumed
using the EADP constraint (SHELX-97) for both M2ϩ ions and
restraining the sum for both positions to 1. Refinement of the iron
compound without this restraint gave no hint on an occupancy
lower than one, as a possible result of the presence of FeIII. Ad-
ditionally, disorder of Ti4ϩ among the two general positions Ti1
Fig. 4 Correlation between an orthorhombic F-centred unit cell
(Fddd, a ϳ 7.35 A, b ϳ 7.35 A, c ϳ 12.5 A) and the monoclinic
cell of the oxidephosphates MTi2O2(PO4)2 (M: Fe, Co, Ni; P21/c,
˚
˚
˚
˚
˚
˚
a ϳ 7.35 A, b ϳ 7.35 A, c ϳ 7.35 A, β ϳ 120°). The orthorhombic
cell is related the unit cell of tetragonal β-Fe2O(PO4) (I41/amd,
˚
˚
a ϭ 5.344(5) A, c ϭ 12.460(8) A) by the transformation matrix
(1 Ϫ1 0 1 1 0 0 0 1) (Fig. 1).
Z. Anorg. Allg. Chem. 2007, 2568Ϫ2578
 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.zaac.wiley-vch.de
2571

M. Schöneborn, R. Glaum
and Ti2 (in one octahedral void) was introduced into the structure
models. The resulting site occupation factors for MII and TiIV re-
fined independently to almost identical values, as expected from
the disorder model (see “Discussion“). All refinements converged
at reasonable residuals and crystal chemical parameters.
tropic displacement parameters and interatomic distances are given
in Tables 6 and 7. Supplementary material has been deposited with
Fachinformationszentrum Karlsruhe, Abt. IDNT, D-76344 Eggen-
stein-Leopoldshafen (e-mail: crysdata@fiz-karlsruhe.de) and can
be obtained by contacting FIZ [reference numbers: CSD-418196
(FeTi2O2(PO4)2), CSD-418197 (CoTi2O2(PO4)2), CSD-418198
(NiTi2O2(PO4)2)].
Information on crystal data, intensity measurement and structure
refinement is summarized in Table 5. Final atomic coordinates, iso-
Angular Overlap Modeling of MTi2O2(PO4)2
(M ؍ Fe, Co, Ni)
Oxidephosphates MTi2O2(PO4)2 (M ϭ Fe, Co, Ni) are con-
taining trimeric units [MIITiIV2O12] of three face-sharing
octahedra (e. g. [FeIITiIV2O12] in Fig. 6). For the chromo-
phores [MIIO6] strong trigonal distortions with angles
Є(O,MII,O) ϳ 79° and 101° are observed. For a better
understanding of the correlation between angular distortion
of the chromophores and their d-electron energies calcu-
lations within the framework of the angular overlap model
(AOM) [38, 28, 29] were performed. An advantage of this
model is its ability to use the chromophores with their ac-
tual geometry, as determined from crystal structure analy-
sis. Instead of using global parameters, like 10Dq or ∆o,
one σ- and two π-interactions (in total 18 bonding param-
eters for an octahedral chromophore) of each ligand with
the five 3d-orbitals of the central ion are used for the fitting
Fig.
β-Fe2O(PO4)
MTi2O2(PO4)2 (M: Fe, Co, Ni) (c, d).
5
Comparison of the crystal structures of tetragonal
(a, b) and monoclinic oxidephosphates
Table 5 Crystal data for MTi2O2(PO4)2 (M: Fe, Co, Ni).
Chemical Formula
FeTi2O2(PO4)2
CoTi2O2(PO4)2
NiTi2O2(PO4)2
Formula weight / g·molϪ1
Crystal system
Space group
373.59
monoclinic
P21/c (No. 14)
376.67
monoclinic
P21/c (No. 14)
376.45
monoclinic
P21/c (No. 14)
Lattice parameters (from Guinier photographs)
˚
a/ A
7.394(3)
7.396(2)
7.401(3)
120.20(3)
349.8(2)
2
4.777
3.547
dark red
prismatic
0.25·0.17·0.14
360
7.381(6)
7.371(5)
7.366(6)
120.26(6)
346.2(5)
2
5.128
3.614
pinkish red
square disc
0.28·0.25·0.14
362
7.388(4)
7.334(10)
7.340(3)
120.12(4)
344.0(2)
2
5.486
3.635
˚
b/ A
˚
c/ A
b/ °
3
˚
Unit cell volume / A
Z
Absorp. coeff. / mmϪ1
Dx-ray / g·cm3
Crystal colour
Crystal shape
Crystal size / mm3
F(000)
green
irregular-shaped isometric
0.17·0.125·0.12
364
Temperature [K]
293(2)
293(2)
293(2)
˚
Wavelength (Mo-Kα, 0.71073 A, graphite monochromator, κ-CCD diffractometer (Nonius)
Theta range / °
Absorption correction
3.18 Յ θ Յ 39.95
multiscans,
within data reduction
using “Scalepack” [67]
4195
2160 [R(int) ϭ 0.020]
Ϫ13 Յ h Յ 13
Ϫ13 Յ k Յ 13
Ϫ13 Յ l Յ 13
75
3.18 Յ θ Յ 30.04
multiscans,
within data reduction
using “Scalepack” [67]
4010
1059 [R(int) ϭ0.037]
Ϫ10 Յ h Յ 10
Ϫ10 Յ k Յ 10
Ϫ10 Յ l Յ 10
76
3.18 Յ θ Յ 35.05
Multiscans [68]
after data reduction in
WinGx [26]
Reflections collected
Independent reflections
Index ranges
14618
1564[R(int) ϭ 0.053]
Ϫ11 Յ h Յ 11
Ϫ11 Յ k Յ 11
Ϫ11 Յ l Յ 11
77
(„whole sphere“)
Parameters
R indices [I > 2σ(I)]
R1 ϭ 0.039
wR2 ϭ 0.116
R1 ϭ 0.042
R1 ϭ 0.031
wR2 ϭ 0.074
R1 ϭ 0.032
R1 ϭ 0.054
wR2 ϭ 0.138
R1 ϭ 0.056
R indices (all data)
wR2 ϭ 0.117
1.301
single-crystal
A ϭ 0.0398, Bϭ 1.5798
Ϫ
wR2 ϭ 0.075
1.224
0.046(1)
A ϭ 0.0056, B ϭ 2.2109
Ϫ
0.848 and Ϫ0.798
wR2 ϭ 0.143
1.235
0.472(3)
A ϭ 0.0001, Bϭ 8.2246
0.002(2)
2.183 and Ϫ1.460
Goodness-of-fit on F2
BASF
Weighting scheme
Extinction coefficient
Ϫ3
˚
Largest diff. peak and hole / e·A
1.878 and Ϫ1.377
2572
www.zaac.wiley-vch.de
 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2007, 2568Ϫ2578

Lazulite-Type Oxidephosphates MTi2O2(PO4)2 (M ϭ FeII, CoII, NiII)
Table
6 Atomic coordinates und isotropic displacement
parameters of a) NiTi2O2(PO4)2 (Rietveld-Refinement [2]),
b) NiTi2O2(PO4)2 (single-crystal refinement, this work), c)
CoTi2O2(PO4)2, d) FeTi2O2(PO4)2.
2
a)
˚
Atom
Position
x
y
z
s.o.f
Ueq [A ]
M1
a
b
c
d
a
b
c
d
a
b
c
d
a
b
c
d
a
b
c
d
a
b
c
d
a
b
c
d
a
b
c
2a
2a
2a
2a
2d
2d
2d
2d
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
4e
0
0
0
0
Ϫ
1/2
1/2
1/2
0.7348(5)
0.7352(2)
0.7334(1)
0.7340(1)
Ϫ
0
0
0
0
Ϫ
0
0
0
0
0
0
0
Ϫ
1/2
1/2
1/2
0.5301(5)
0.5291(2)
0.5296(1)
0.5290(1)
Ϫ
1
0.005
0.0067(3)
0.898(3)
0.9799(19) 0.0071(2)
0.9724(16) 0.0079(1)
M2b)
Ti1
Ti2b)
P
Ϫ
Ϫ
0.102(3)
0.0201(19) 0.0071(2)
0.0276(16) 0.0079(1)
0.0067(3)
0.2203(5)
0.2215(2)
0.2215(1)
0.2227(1)
Ϫ
1
0.006
0.0064(3)
0.912(4)
0.9799(24) 0.0055(2)
0.9753(16) 0.0063(1)
Ϫ
Ϫ
0.7650(2)
0.2780(2)
0.4820(2)
0.088(4)
0.0064(3)
0.7567(48) 0.2774(50) 0.4741(51) 0.0201(24) 0.0055(2)
0.7640(30) 0.2740(20) 0.4780(30) 0.0247(16) 0.0063(1)
0.2462(9)
0.2464(2)
0.2458(1)
0.2471(1)
0.7640(2)
0.7640(9)
0.7623(4)
0.7612(3)
0.7766(2)
0.7863(8)
0.7823(4)
0.7793(3)
0.4400(2)
0.4426(8)
0.4399(4)
0.4404(3)
0.2579(2)
0.2673(9)
0.2615(4)
0.1614(3)
0.0578(2)
0.0564(8)
0.0543(4)
0.0560(3)
0.1237(9)
0.1238(2)
0.1224(1)
0.1213(1)
0.1589(2)
0.1528(6)
0.1559(4)
0.1574(3)
0.0031(2)
0.0013(8)
0.0051(4)
0.0049(2)
0.2415(2)
0.2389(8)
0.2392(4)
0.2390(3)
0.0086(2)
0.0099(8)
0.0097(4)
0.0087(3)
0.2481(2)
0.2491(8)
0.2515(4)
0.2530(2)
0.7508(8)
0.7502(2)
0.7488(1)
0.7483(1)
0.7671(2)
0.7648(7)
0.7663(4)
0.7634(3)
0.0939(2)
0.1003(7)
0.1004(4)
0.1002(3)
0.8716(2)
0.8745(8)
0.8695(4)
0.8669(3)
0.5816(2)
0.5883(8)
0.5834(4)
0.5835(3)
0.1452(2)
0.1423(8)
0.1438(4)
0.1444(3)
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0.008
0.0041(3)
0.0044(2)
0.0055(1)
0.003
0.0113(9)
0.0089(5)
0.0103(3)
0.005
0.0106(8)
0.0090(4)
0.0097(3)
0.003
0.0129(9)
0.0103(5)
0.0109(3)
0.005
0.0147(10)
0.0108(5)
0.0113(3)
0.004
0.0142(9)
0.0087(5)
0.0093(3)
O1
O2
O3
d
a
b
c
d
a
b
c
O4
O5
d
a)
U
ϭ (1/3)ΣiΣjUijai*aj*ai ·aj
Fig. 6 Trimer [FeIITiIV2O12]. Ellipsoids are given at the 90 % pro-
bability level.
eq
b) In the refinement of NiTi2O2(PO4)2 from powder data [2] no disorder of the metal
cations was allowed for.
between calculated and observed transition energies. This
decomposition of the global ligand field parameter (10Dq
or ∆o) into parameters describing individual bonding inter-
actions permits also accounting for second-sphere ligand
field effects [30], e. g. cation-cation interaction. To reduce
the number of independent bonding parameters, constraints
between some of them were introduced. Thus, for the en-
ergy eσ(MII-O), proportionality to the distance d(M-O)Ϫ5.0
is assumed [38, 31]. In general, the energy of eπ is set to
one quarter of the corresponding energy eσ in the case of
“undisturbed” π-interaction [15, 38]. In the case of the par-
ticular bonding situation encountered in the oxidephos-
phates eπ ϳ 0 has been assumed. The validity of this as-
sumption will be discussed.
Interelectronic repulsion is introduced into the AOM cal-
culations via the Racah parameters B and C, spin orbit
coupling by ξ. For the angular overlap modelling the free
ion ratios C0/B0 ϭ 4.3 (FeII), 3.9 (CoII), and 4.3 (NiII) were
used and retained during the calculations [38]. Variations in
the covalency of the M-O interaction in the complexes were
considered by variable nephelauxetic ratios β [β ϭ B/B0;
B0 ϭ 897 cmϪ1 (FeII), 989 cmϪ1 (CoII), 1042 (NiII)] [38].
Table 7 MTi2O2(PO4)2 (M: Fe, Co, Ni). Interatomic distances /
A in trimers [M Ti2O12] and in [PO4] groups. Distances are given
only for the mainly occupied sites M1 and Ti1.
II
˚
FeTi2O2(PO4)2
[Fe1O6]
CoTi2O2(PO4)2
[Co1O6]
NiTi2O2(PO4)2
[Ni1O6]
Fe1-O1 (2x) 2.105(2) Co-O1 (2x) 2.078(3) Ni1-O1 (2x) 2.063(5)
Fe1-O2 (2x) 2.105(2) Co-O2 (2x) 2.080(3) Ni1-O2 (2x) 2.050(5)
Fe1-O5 (2x) 2.091(2) Co-O5 (2x) 2.072(3) Ni1-O5 (2x) 2.041(5)
[Ti1O6]
Ti1-O1
Ti1-O1
Ti1-O2
Ti1-O3
Ti1-O4
Ti1-O5
Fe1-Ti1
[PO4]
P1-O2
P1-O3
P1-O4
P1-O5
[Ti1O6]
1.713(2) Ti1-O1
2.255(2) Ti1-O1
2.067(2) Ti1-O2
1.901(2) Ti1-O3
1.910(2) Ti1-O4
2.094(2) Ti1-O5
2.923(1) Co1-Ti1
[PO4]
1.545(2) P1-O2
1.519(2) P1-O3
1.524(2) P1-O4
1.536(2) P1-O5
[Ti1O6]
1.717(3) Ti1-O1
2.245(3) Ti1-O1
2.067(3) Ti1-O2
1.895(3) Ti1-O3
1.905(3) Ti1-O4
2.086(3) Ti1-O5
2.927(2) Ni1-Ti1
[PO4]
1.546(3) P1-O2
1.517(3) P1-O3
1.526(3) P1-O4
1.535(3) P1-O5
1.709(5)
2.253(5)
2.086(6)
1.893(5)
1.899(6)
2.094(6)
2.909(2)
1.540(5)
1.522(5)
1.523(5)
1.533(5)
The spin-orbit coupling parameter ξ was also assumed to
be reduced relative to the free ion values (ξ0(FeII) ϭ
Z. Anorg. Allg. Chem. 2007, 2568Ϫ2578
 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.zaac.wiley-vch.de
2573

M. Schöneborn, R. Glaum
400 cmϪ1, ξ0(CoII) ϭ 515 cmϪ1, ξ0(NiII) ϭ 630 cmϪ1
according to β.
For the AOM calculations the PC program CAMMAG
[32, 33] in a modified version [34] was used. Best fit AOM
parameters for the spectra of NiTi2O2(PO4)2 and
CoTi2O2(PO4)2 are summarised in Table 8.
)
spectrum. Neglecting eσ (Co-Ti) allows also a fit to the ob-
served band energies, which was, however, less good.
Angular overlap modelling on the basis of the bonding
model derived for the nickel(II) and cobalt(II) oxidephos-
phate did not allow a reasonable fit to the absorption bands
observed for FeTi2O2(PO4)2. In particular the rather large
band splittings due to low-symmetry ligand-fields were not
reproduced by the calculations. We attribute this failure to
the presence of a dynamic Jahn-Teller effect (see Dis-
cussion).
Table 8 Summary of ligand-field parameters for FeII, CoII, and
NiII in oxidephosphates MTi2O2(PO4)2.
Fe
Co
Ni
Racah parameters for the free ions [38]
Results and Discussion
Chemical vapour transport
B0 / cm؊1
C0 / cm؊1
C/B
897
3857
4.3
989
4253
4.3
1042
4585
4.4
ξ0 / cm؊1
400
515
630
The CVT experiments allow crystallization of oxidephos-
phates MTi2O2(PO4)2 (MII: Fe, Co, Ni), although the ob-
served deposition rates are rather low (a few mg/d). Mi-
gration of these phosphates proceeds always from the
higher to the lower temperature (T2 Ǟ T1) due to endo-
thermic reactions. Gaseous mixtures of Cl2, PCl3, and TiCl4
(molar ratio 0.26 : 0.03 : 0.03; from in situ reaction of TiP
with a surplus of PtCl2) are in general suitable as transport
agent for the oxidephosphates, even though such mixtures
do also attack the wall of the silica tube. For NiTi2O2(PO4)2
no other transport agent allowed controlled crystallisation.
For CoTi2O2(PO4)2 plain chlorine, too, used as transport
agent, leads to a migration of the oxidephosphate via the
gas phase. Transport experiments aiming at the crystallis-
ation of CoTi2O2(PO4)2 or FeTi2O2(PO4)2 led under vari-
ous experimental conditions to formation and chemical
vapour transport of NASICON [35, 36] type phosphates
MIITi4(PO4)6 [M: Fe (orange-red), Co (lavender colour)].
This is apparently one of the very rare cases, where crystals
of NASICON type compounds suitable for a single-crystal
structure investigation were obtained. Detailed crystallo-
graphic investigations on these phosphates are in progress
[1].
Tanabe-Sugano evaluation
Bo/ cm؊1
a)
∆
/ cm؊1
9660
Ϫ
Ϫ
8030
895
0.90
7800
925
0.89
β ؍ B/B0
b)
Angular overlap modelling
e
e
σ,max(M-O) / cm؊1
Ϫ
Ϫ
Ϫ
Ϫ
Ϫ
2875
3620
Ϫ500
811
2850
3420
Ϫ500
854
c)
σ,norm(M-O)2.0 / cm؊1
eσ(M-Ti) / cm؊1
B / cm؊1
β ؍ B/B0
0.82
0.82
a) Calculated from the barycentre of the band around 9900 cmϪ1, assuming
5
for the T2g ground state a splitting, which is one quarter of the observed
splitting of the excited 5Eg state.
b) σ-interactions between M and oxygen were chosen according to previous
investigations [17, 40, 41], they were assumed to be proportional to
d(M-O)Ϫ5.0. π-interactions were neglected (eπ ϭ 0 cmϪ1), see “Discussion“;
ξ was reduced relative to ξ0 as B to B0, the ratios C/B of the free ions were
kept during the modelling.
c)
˚
σ-interaction between M and oxygen normalized to d(M-O) ϭ 2.00 A,
using the relation eσ ϳ d(M-O)Ϫ5.0 [31].
As a first step the transition energies in the spectrum of
NiTi2O2(PO4)2 were modelled. Using eσ,max(Ni-O) ϭ
3000 cmϪ1 (for the shortest observed, Tab. 7, distance
d(Ni-Oi)Ϫ5.0),
˚
d(Ni-O)
ϭ
2.041 A and eσ(Ni-Oi)
ϳ
eπ(Ni-O) ϭ 0 cmϪ1, and B ϭ 875 cmϪ1 (C/B ϭ 4.4, β ϭ
Using plain chlorine (from in situ decomposition of
0.84, ξ ϭ 529 cmϪ1) allowed a reasonable fit (Fig. 3a) to
PtCl2) for CVT of CoTi2O2(PO4)2, CoCl2,g, TiCl4,g, O2,g
,
˜
˜
˜
bands ν1 and ν2. Even the rather large splitting of ν2 was
and P4O10,g are expected as volatile species (eq. (3)). The
dominance of gaseous P4O10 as phosphorus carrier under
such conditions has already been discussed [15, 16]. The
corresponding equilibrium with NiTi2O2(PO4)2 is appar-
ently thermodynamically less favourable and does not allow
volatilization of the phosphate. FeTi2O2(PO4)2, on the other
hand, is oxidized by chlorine in a quantitative reaction (eq.
˜
reproduced. For ν3, however, this calculation leads to a
splitting of ca. 2500 cmϪ1, which is not in accordance with
the observed spectrum (Fig. 3). Searching for an improved
match differences in the σ-interaction of NiII with oxide and
oxygen of a phosphate group were considered. This clearly
changed the fit, especially in the magnitude of the band
splittings, to the worse (Fig 3b). Eventually, weak interac-
tions with σ-symmetry (eσ ϭ Ϫ500 cmϪ1) between NiII and
adjacent TiIV were assumed. In other words, the TiIV cat-
ions were considered as additional ligands of NiII. Doing
so led to a nice match of all observed spectral features
(eσ,max(Ni-O) ϭ 2850 cmϪ1, B ϭ 854 cmϪ1, C/B ϭ 4.4, β ϭ
0.82, ξ ϭ 517 cmϪ1; Fig. 3c).
(4)), leading to the formation of FeIII4TiIV27O24(PO4)24,s
Fe2O3,s, FePO4,s, and FeCl3,g
,
.
CoTi2O2(PO4)2,s ϩ 5 Cl2,g ϭ CoCl2,g ϩ 2 TiCl4,g
ϩ 1/2 P4O10,g ϩ 5/2 O2,g (3)
27 FeTi2O2(PO4)2,s ϩ 27/2 Cl2,g ϭ 2 Fe4Ti27O24(PO4)24,s
ϩ 2 Fe2O3,s ϩ 6 FePO4,s ϩ 9 FeCl3,g (4)
Using for CoTi2O2(PO4)2 the same bonding model
(eσ,max(Co-O) ϭ 2875 cmϪ1, eσ (Co-Ti) ϭ Ϫ500 cmϪ1, B ϭ
811 cmϪ1, C/B ϭ 4.3, β ϭ 0.82, ξ ϭ 422 cmϪ1; Fig. 3d)
immediately gave a very good match with the observed
Addition of a few mg TiP to chlorine prevents oxidation
of iron(II)-oxidephosphate and creates favourable transport
conditions (low oxygen equilibrium pressures) for all oxide-
2574
www.zaac.wiley-vch.de
 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2007, 2568Ϫ2578

Lazulite-Type Oxidephosphates MTi2O2(PO4)2 (M ϭ FeII, CoII, NiII)
phosphates. Mixtures Cl2,g/PCl3,g (from oxidation of TiP by
chlorine) are strongly chlorinating reagents. Their applica-
bility for chemical vapour transport of lanthanum mono-
phosphate LaPO4 has been pointed out by Schäfer, almost
40 years ago [37]. Eq. (5) is likely to be the dominating
transport reaction under these conditions.
[MIITiIV2O12] (Fig. 6). In the real structure the empty sites
“ᮀ” are to a small extend (up to 10 %) occupied by MII
(site M2, Wyckoff position 2d), as is revealed by our single
crystal X-ray investigations (Table 6). The disorder of the
MII cations causes in their immediate vicinity a relaxation
of the lattice. Titanium(IV) is shifted due to the electronic
repulsion between MII and TiIV to split position Ti2. An-
other consequence of the electrostatic repulsion between
MII and TiIV cations within the trimers [MIITi2O12] are
strong trigonal distortions of the coordination polyhedra
[MIIO6] (point group D3d) and [TiIVO6] (point group C3v)
(Tables 7; Fig. 6). In contrast to β-Fe2O(PO4) the coordi-
nation of the oxide ions is reduced to c.n.(O) ϭ 3, oxygen
atoms of phosphate groups are connected to one (O3, O4)
or two metal cations (O2, O5).
Group theoretical considerations of the β-Fe2O(PO4) and
MTi2O2(PO4)2 structures elucidate their close relation. The
lattice parameters and the atomic coordinates derived from
the aristotype β-Fe2O(PO4) for the MTi2O2(PO4)2 struc-
tures by group-subgroup considerations are consistent with
those obtained from the single crystal X-ray investigations
(Fig. 7). The translationsgleiche transition of index 2 from
space group Fddd (crystal class 2/m 2/m 2/m) to C2/c
(crystal class 2/m) means loss of a two-fold axis parallel
the b-axis in the lattice of the supergroup. This symmetry
operation is conserved as the twinning element, meaning a
two-fold twin-axis along the a-axis of the subgroup. The
klassengleiche transition of index 2 from space group C2/c
to P21/c leads to a splitting of the Wyckhoff position 4c
into two different positions 2a and 2d with the same site
symmetry and the same chemical environment. Such a situ-
ation causes frequently disorder, as it is indeed observed for
cations MII.
MTi2O2(PO4)2,s ϩ 2 Cl2,g ϩ 2 PCl3,g ϭ MCl2,g
ϩ 2 TiCl4,g ϩ P4O10,g (M: Fe, Co, Ni) (5)
It is worth noting that most transport experiments with
FeTi2O2(PO4)2 and CoTi2O2(PO4)2 led to the formation of
additional phosphates [M2P2O7, MTi4(PO4)6; M: Fe, Co;
Table 3 and 4] in the equilibrium solids. This observation
points to an incongruent gasphase solubility for FeTi2-
O2(PO4)2 and CoTi2O2(PO4)2, with an accumulation of ti-
tanium as tetrachloride in the gas phase, while MO and
P4O10 are enriched in the solid phase (eq. (6)). Conse-
quently, a small surplus of TiO2 in the starting mixture
avoids formation of the neighbouring phosphates (Table 3).
5 MTi2O2(PO4)2,s ϩ 12 Cl2,g ϭ 2 M2P2O7,s ϩ MTi4(PO4)6,s
ϩ 6 TiCl4,g ϩ 6 O2,g (M: Fe, Co) (6)
In our experiments the new iron(III)-oxidephosphate
FeIII4TiIV27O24(PO4)24,s (isotypic to MIII4TiIV27O24(PO4)24,s
(M: Ti, Cr) [54, 56]) and NASICON type phosphates
FeIITiIV4(PO4)6 and CoIITiIV4(PO4)6 were obtained for the
first time as single crystals suitable for detailed structural
studies. Such investigations are in progress [1].
Crystal Structure
The oxidephosphates MTi2O2(PO4)2 (MII: Fe, Co, Ni) are
members of the Lipscombite/Lazulite structure family (Fig.
4, 5). Their crystal structure can be derived from that of
tetragonal β-Fe2O(PO4), which is the aristotype of this
structure family. β-Fe2O(PO4) crystallizes in space group
I41/amd (a ϭ 5.344(5) A, c ϭ 12.460(8) A, Z ϭ 4) [10, 11].
Its structure is built up by chains of face-sharing octahedra
[FeII/IIIO6] along the [1 0 0] and [0 1 0] directions (Fig. 5a
and b). Each octahedron shares four vertices with [PO4]
tetrahedra. The two remaining vertices are occupied by ox-
ide ions, which crosslink the chains. The oxide ions are co-
ordinated by four iron cations while the oxygen atoms of
phosphate groups are connected to two iron cations (and
phosphorus; c.n.(O) ϭ 3; Fig. 4).
Figs. 4 and 5 illustrate the relationship between the struc-
tures of β-Fe2O(PO4) and MTi2O2(PO4)2. In β-Fe2O(PO4)
all octahedral voids are occupied by iron. Substitution of
FeII/III by MII/TiIV leads to an ordered introduction of vac-
ancies into the chains of face-sharing octahedra, which is
accompanied by a reduction of symmetry from space group
Colour and electronic absorption spectra
Crystals of FeTi2O2(PO4)2 are deeply red (Fig. 2a). The ab-
sorption spectrum (Fig. 3a) shows a broad band in the nir
region with two maxima at 8700 cmϪ1 and 11000 cmϪ1. The
red color results from rather low-lying charge-transfer tran-
sitions, that block all light above 16500 cmϪ1. Crystals of
CoTi2O2(PO4)2 are light red with a pinkish tint (Fig. 2b).
˚
˚
Three broad absorption bands at ν1 ϭ 8100 cmϪ1, ν2
˜
˜
ϭ
15600 cmϪ1, and ν3 ϭ 19500 cmϪ1 (with a maximum at
18900 cmϪ1 and a shoulder at 20400 cmϪ1) are observed
(Fig 3b). Evaluation of the spectrum using the Tanabe-
Sugano diagram in ref. [38] led for the chromophore
[CoIIO6] to the ligand-field splitting ∆o ϭ 8030 cmϪ1 and
the Racah parameter B ϭ 895 cmϪ1 (β ϭ B/Bfree ion ϭ 0.90).
Crystals of NiTi2O2(PO4)2 are light green (Fig. 2c). Three
˜
strong absorption bands at ν1 ϭ 8000 cmϪ1, ν2
ϭ
˜
˜
13400 cmϪ1, and ν3 ϭ 24000 cmϪ1 are found (Fig. 3c). The
band around 13400 cmϪ1 appears to be very broad, poss-
ibly with three components (a, b, c) at 12000, 13400 and
14600 cmϪ1. The band around 24000 cmϪ1 shows a maxi-
mum at 24700 cmϪ1 and a shoulder at 23600 cmϪ1, but is
still rather narrow. At about 21000 cmϪ1 a weak band,
˜
˚
˚
˚
I41/amd to P21/ c (a ϳ 7.3 A, b ϳ 7.3 A, c ϳ 7.4 A,
β ϳ 120°, Z ϭ 2). In the ideal structure of MTi2O2(PO4)2
the occupancy sequence for the octahedral voids is ..., Ti4ϩ
,
MII, TiIV, ᮀ, ... thus the main structural motif are trimers
Z. Anorg. Allg. Chem. 2007, 2568Ϫ2578
 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.zaac.wiley-vch.de
2575

M. Schöneborn, R. Glaum
Fig. 7 Group-subgroup relation between the crystal structures of β-Fe2O(PO4) and FeTi2O2(PO4)2.
possibly originating from a spin-forbidden electronic tran-
sition is observed. The observed spectrum is very similar,
however better resolved than the one reported in [2]. Based
on the baricentres of the three broad bands at 8000, 13500,
and 24000 cmϪ1 an evaluation according to Tanabe and Su-
gano leads for the chromophore [NiIIO6] to ∆o ϭ 7832 cmϪ1
and B ϭ 926 cmϪ1 (β ϭ B/Bfree ion ϭ 0.89).
The colours of oxide-phosphates MTi2O2(PO4)2 (M ϭ Fe
(deep-red), Co (pinkish-red), Ni (green)) differ considerably
from those of other anhydrous phosphates containing these
cations (iron(II) phosphates: colourless [39], cobalt(II)
phosphates: purple to violett colour [40], nickel(II) phos-
phates: orange to yellow [17, 41, 42]). In the case of
FeTi2O2(PO4)2 the deep-red colour clearly is not the result
of d-d electronic transitions, which are observed in the near-
infrared region. Apparently, the colour originates from
some sort of charge-transfer transition. Comparison with
CoTi2O2(PO4)2 and NiTi2O2(PO4)2 suggests that
LMCT(OϪII Ǟ TiIV) transitions should be at higher energy
than the observed 16500 cmϪ1. This points to the possibil-
ity of IVCT transitions (FeII Ǟ TiIV and/or FeII Ǟ FeIII).
In comparison to other anhydrous nickel(II) phosphates
(e. g. NiP4O11 and CaNiP2O7 [41]) and many chromophores
containing Ni2ϩ ions in undistorted octahedral environ-
ment, the spectrum of NiTi2O2(PO4)2 shows three
distinct differences. Firstly, the ligand-field splitting
∆(NiTi2O2(PO4)2) ϭ 7800 cmϪ1 is roughly 1000 cmϪ1
higher than the splittings observed for NiP4O11 or
CaNiP2O7. The splitting (ϳ2500 cmϪ1) of the second spin-
˜
allowed transition (ν2 in Fig. 3) is significantly larger and,
thirdly, the interelectronic repulsion parameters are smaller
than in other anhydrous nickel(II) phosphates. As shown in
Fig. 3c angular overlap modelling allows a good fit to these
pecularities. In addition, the same bonding model gives
with almost identical parameters a perfect match to the
spectrum observed for CoTi2O2(PO4)2 (Fig. 3d). The ob-
tained best-fit AOM parameters are in good agreement with
parameters derived for large series of nickel and cobalt
phosphates [17, 40]. According to the AO modelling the
particular bonding situation of NiII (and CoII) in the tri-
mers [MIITiIV2O12] is characterized by several geometric
and electronic consequences of the close vicinity between
2576
www.zaac.wiley-vch.de
 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2007, 2568Ϫ2578

Lazulite-Type Oxidephosphates MTi2O2(PO4)2 (M ϭ FeII, CoII, NiII)
TiIV and MII. a) Face-sharing of octahedra [TiIVO6] and
[MIIO6] leads to strong radial and angular distortion of the
latter. b) Ti4ϩ (d0 system) is strongly π-accepting, thus leav-
ing no electron density at the bridging oxygen atoms for π-
interaction with the d-orbitals of the nickel ions. The exclus-
ive presence of σ-donating oxygen ligands without any π-
donating effect explains the rather large ligand-field split-
ting. Similar observations have been reported for NiII-con-
taining perowskites (e. g. Sr2NiIIMVIO6, MVI: W, Te), where
d0-cations in the vicinity of NiII lead to a dramatically
higher ligand-field splitting in the [NiIIO6] chromophores
than d10-cations like TeVI [43]. c) The angular distortion of
the chromophore [NiIIO6] together with the weak interac-
tion between TiIV and NiII accounts for the large splitting
of the second absorption band. d) This same interaction
allows also weak delocalization of d-electron density from
NiII to TiIV, similar to the situation already described for
CrIII doped into Ga2O3 or Al2O3 [44].
Obviously, the rather simple AOM approach allows a
nice description of the fairly complex bonding situation
encountered by NiII and CoII in the oxidephosphates
MTi2O2(PO4)2 (M: Co, Ni). Even second-sphere ligand
field effects [30] like direct MII-TiIV interaction and reduced
π-bonding are well accounted for. Considering the detailed
understanding for the bonding situation of NiII and CoII in
the oxidephosphates, it was quite disappointing, however,
not unexpected, that the d-d electronic transitions observed
amples are known in literature for vibronic coupling in FeII
and CuII containing chromophores as an explanation for
largely split absorption bands, despite highly symmetrical
2ϩ
coordination polyhedra (e. g. Fe(H2O)6 [46], KFeF3 [47],
2ϩ
4Ϫ
Cu(H2O)6
[48], CuF6
[49]). With eσ,max(Fe-O) ϭ
3400 cmϪ1 at dmin(Fe-O) ϭ 2.090 A (Tab. 7) for the
dynamically undistorted chromophore and eσ,eq(Fe-O) ϭ
3735 cmϪ1 and eσ,ax(Fe-O) ϭ 2470 cmϪ1 for the dynami-
cally distorted chromophore one may estimate using the re-
lation eσ(FeII-O) ϳ d(Fe-O)Ϫ5.0 a dynamically elongated
˚
˚
distortion with deq(Fe-O) ϭ 2.05 A and dax(Fe-O) ϭ
2.22 A. These distances obtained for the vibronically dis-
˚
torted chromophore [FeIIO6] in FeTi2O2(PO4)2 coincide
nicely with those observed for statically distorted polyhedra
[FeIIO6] in FeP4O11 [50], Fe2P4O12 [51], and Fe3(PO4)2 [52].
Our considerations do not allow to distinguish between vib-
ronic tetragonal elongation and compression nor do they
provide exclusive evidence for the tetragonal distortion
along the direction iron(II)Ϫoxide. However, we believe
that such a vibronic distortion is the most reasonable.
Acknowledgment. We thank Dr. Jörg Daniels (Bonn University) for
collecting the single-crystal diffraction data sets and Prof. Dr.
Johannes Beck for the friendly provision of the instrument. We also
thank one reviewer for valuable comments and suggestions.
References
˜
for FeTi2O2(PO4)2 (Fig. 3) and α-CuTi2O2(PO4)2 (ν1
ϭ
[1] M. Schöneborn, part of planned Ph. D. thesis, Univ. of Bonn.
[2] P. Graverau, J. P. Chaminade, B. Manoun, S. Krimi, A. El
Jazouli, Powder Diff. 1999, 14, 10.
7940 cmϪ1, ν2 ϭ 10240 cmϪ1, and ν3 ϭ 12640 cmϪ1 [4])
could not be matched by calculations within the AOM
framework.
˜
˜
[3] S. Benmokhtar, A. El Jazouli, J. P. Chaminade, P. Gravereau,
`
A. Wattiaux, L. Fournes, J. C. Grenier, D. Waal, J. Solid State
Allowing within the AOM framework for the static geo-
metric distortions (radial and angular) of the chromophore
[FeIIO6] and introduction of second-sphere ligand field ef-
fects reduces the symmetry of the ligand field from Oh to
D3d. According to group theory, such a symmetry reduction
Chem. 2006, 179, 3709.
[4] S. Benmokhtar, H. Belmal, A. El Jazouli, J. P. Chaminade, P.
Gravereau, S. Pechev, J. C. Grenier, G. Villeneuve, D. de Waal,
J. Solid State Chem. 2007, 180, 772.
[5] P. Graverau, S. Benmokhtar, J.-P. Chaminade, A. El Jazouli,
E. Lebraud, D. Denux, Solid State Sci. 2007, 9, 258.
[6] L. Katz, W. N. Lipscomb, Acta Crystallogr. 1951, 4, 345.
[7] M. A. Gheith, Amer. Miner. 1953, 38, 612.
[8] H. Bärnighausen, MATCH Ϫ Commun. Math. Chem. 1980,
9, 139.
[9] H. D. Megaw, “Crystal structures: A working approach“,
Saunders Co., Philadelphia, 1973.
[10] M. Ijaali, B. Malaman, C. Gleitzer, J. K. Warner, J. A. Hriljac
and A. K. Cheetham, J. Solid State Chem. 1990, 86, 195.
[11] B. Ech-Chahed, F. Jeannot, B. Malaman, C. Gleitzer, J. Solid
State Chem. 1988, 74, 47.
[12] E. Elkaim, J. F. Berar, C. Gleitzer, B. Malaman, M. Ijjaali, C.
Lecomte, Acta Crystallogr. 1996, B52, 428.
[13] J. L. Pizarro, G. Villeneuve, P. Hagenmuller, A. Le Bail, J.
Solid State Chem. 1991, 92, 273.
[14] H. Schäfer, Chemical Transport Reactions, Academic Press,
New York, 1964.
[15] R. Glaum, Neue Untersuchungen an wasserfreien Phosphaten
der Übergangsmetalle (in german), Thesis of Habilitation,
Univ. of Gießen, 1999. URL: http://bibd.uni-giessen.de/ghtm/
1999/uni/h990001.htm
5
does not cause a splitting of the excited electronic Eg state
for iron(II). Results of AOM calculations [eσ,max(Fe-O) ϭ
3400 cmϪ1, B ϭ 736 cmϪ1, C/B ϭ 4.3, β ϭ 0.82, ξ ϭ
328 cmϪ1, parameters in analogy to the nickel(II) and co-
balt(II) oxidephosphate] are in agreement with this expec-
tation. They lead to ν( T2g Ǟ Eg) ϭ 9900 cmϪ1, with a
splitting of the 5Eg state of only some 200 cmϪ1, due to
spin-orbit coupling and marginal deviation from the ideal-
ized symmetry. A match of the observed splitting of the
excited 5Eg state (δ ϳ 2400 cmϪ1, Fig. 3) is only obtained in
AOM calculations, if one assumes an additional tetragonal
distortion of the ligand field experienced by the Fe2ϩ ions
5
5
˜
[4 x eσ,eq(Fe-O) ϭ 3735 cmϪ1, 2 x eσ,ax(Fe-O) ϭ 2470 cmϪ1
,
B ϭ 736 cmϪ1, C/B ϭ 4.3, β ϭ 0.82, ξ ϭ 328 cmϪ1]. Such
a distortion as consequence of differences in the bonding
interactions between iron(II) and oxide and iron(II) and
oxygen from a phosphate group, respectively, is in clear con-
tradiction to the modelling of the spectra of NiTi2O2(PO4)2
and CoTi2O2(PO4)2. For d6 and d9 ions like iron(II) and
copper(II), however, such a splitting might be caused by the
presence of a dynamic Jahn-Teller effect [45]. Many ex-
[16] R. Gruehn, R. Glaum, Angew. Chem. 2000, 112, 706; Angew.
Chem. Int. Ed. 2000, 39, 692.
Z. Anorg. Allg. Chem. 2007, 2568Ϫ2578
 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.zaac.wiley-vch.de
2577

M. Schöneborn, R. Glaum
[17] K. Maaß, Ph.D. thesis, Univ. of Gießen, 2002. URL: http://
bibd.uni-giessen.de/ghtm/2002/uni/d020119.htm
[18] A. Winkler, E. Thilo, Z. Anorg. Allg. Chem. 1966, 346, 92.
[19] E. Krausz, AOS News 1998, 12, 21.
[20] E. Krausz, Aust. J. Chem. 1993, 46, 1041.
[21] Y. Ameniya, J. Miyahara, Nature 1998, 336, 89.
[22] I. Tanaka, M. Yao, M. Suzuki, K. Hikichi, J. Appl. Crys-
tallogr. 1990, 23, 334.
[23] K. Maaß, R. Glaum, R. Gruehn, Z. Anorg. Allg. Chem. 2002,
626, 1663.
[24] J. Soose, G. Meyer, “SOS Ϫ Programme zur Auswertung von
Guinier-Aufnahmen“, Univ. of Gießen, 1980.
[41] M. Funke, M. Blum, R. Glaum, B. ElBali, Z. Anorg. Allg.
Chem. 2004, 63, 1040.
[42] G. R. Rossman, R. S. Shannon, R. K. Waring, J. Solid State
Chem. 1981, 39, 277.
[43] D. Reinen, Theoret. Chim. Acta (Berlin) 1966, 5, 312.
[44] D. Reinen, Struct. Bonding (Berlin) 1969, 6, 30.
[45] H. C. Longuet-Higgins, U. Opik, M. H. L. Pryce, R. A. Sack,
Proc. Roy. Soc. (London) 1958, A244, 1.
[46] F. A. Cotton, M. D. Meyers, J. Am. Chem. Soc. 1960, 82, 5023.
[47] G. D. Jones, Phys. Rev. 1967, 155, 259.
[48] M. A. Hitchman, Y. V. Yablokov, V. E. Petrashen, M. A. Au-
gustyniak-Jabokov, H. Stratemeier, M. J. Riley, K. Lukaszew-
icz, P. E. Tomaszewski, A. Pietraszko, Inorg. Chem. 2002, 41,
229.
[49] M. J. Riley, M. A. Hitchman, D. Reinen, Chem. Phys. 1986,
102, 11.
[25] M.
Sheldrick,
“SHELX-97
(Includes
SHELXS97,
SHELXL97, CIFTAB). Programs for Crystal Structure
Analysis (Release 97-2)“, Univ. of Göttingen, Göttingen, Ger-
many, 1998.
[50] M. Weil, R. Glaum, Eur. J. Solid State Inorg. Chem. 1998,
35, 495.
[51] E. A. Genkina, B. A. Maksimov, O. K. Mel’nikov, Kristallog-
[26] L. J. Farrugia, J. Appl. Cryst. 1990, 32, 837.
[27] R. Hundt, “KPLOT“, Version 9.0, Univ. of Bonn 2005.
´
[28] C. K. Jorgensen, R. Pappalardo, H. H. Schmidtke, J. Chem.
rafiya 1985, 30, 885.
Phys. 1963, 39, 1422.
[52] O. V. Yakubovich, E. L. Belokoneva, V. G. Tsirel’son, V. S.
Urusov, Vestnik Mosk. Univ. Geol. 1991, 46, 77.
[53] R. Glaum, R. Gruehn, Z. Kristallogr. 1989, 186, 91.
[54] F. Reinauer, Ph. D. thesis, Univ. Gießen, 1998.
[55] F. Reinauer, R. Glaum, Acta Crystallogr. 1998, B54, 722.
[56] F. Reinauer, R. Glaum, R. Gruehn, Eur. J. Solid State Inorg.
Chem. 1994, 31, 779.
[57] J. T. Vaughey, W. T. A. Harrison, A. J. Jacobson, D. P.
Goshorn, J. W. Johnson, Inorg. Chem. 1994, 33, 2481.
[58] M. L. Lindberg, C. L. Christ, Acta Crystallogr. 1959, 1959, 20.
[59] G. J. Redhammer, G. Tippelt, G. Roth , W. Lottermoser, G.
Amthauer, Phys. Chem. Minerals. 2000, 27, 419.
[60] P. B. Moore, Am. Mineral. 1970, 55, 135.
[61] N. H. W. Sieber, E. Tillmanns, W. Hofmeister, Acta Crys-
tallogr. 1987, C43, 1855.
[29] D. E. Richardson, J. Chem. Ed. 1993, 70, 372.
[30] D. Reinen, M. Atanasov, S. L. Lee, Coord. Chem. Rev. 1998,
175, 91.
[31] M. Bermejo, L. Pueyo, J. Chem. Phys. 1983, 78, 854.
[32] D. A. Cruse, J. E. Davies, J. H. Harding, M. Gerloch, D. J.
Mackey, R. F. McMeeking, “CAMMAG“, a FORTRAN pro-
gramm, Cambridge, 1980.
[33] M. Gerloch, “Magnetism and Ligand Field Theory“, Cam-
bridge Univ. Press, 1983.
[34] M. Riley, “CAMMAG for PC“, V 4.0, Univ. of Queensland,
St.Lucia (Australia), 1997.
[35] H. Kohler, H. Schulz, O. K. Mel’nikov, Mater. Res. Bull. 1983,
18, 1143.
[36] R. Olazcuaga, J. M. Dance, G. Le Flem, J. Derouet, L. Beaury,
P. Porcher, A. El Bouari, A. El Jazouli, J. Solid State Chem.
1999, 143, 224.
[37] H. Schäfer, V. P. Orlovskii, M. Wiemeyer, Z. Anorg. Allg.
Chem. 1972, 390, 13.
[38] B. N. Figgis, M. A. Hitchman, “Ligand Field Theory and Its
Applications“, WILEY-VCH, New York, 2000.
[39] R. Glaum, unpublished results.
[62] C. C. Torardi, W. M. Reiff, L. Takacs, J. Solid State Chem.
1989, 82, 203.
[63] M. Malaman, M. Ijjaali, G. Venturini, C. Gleitzer, J. L. Soub-
eyroux, J. Solid State Chem. 1991, 28, 519.
[64] P. Schmidt-Beurmann, J. Solid State Chem. 2000, 153, 237.
[65] P. Schmidt-Beurmann, J. Mater. Chem. 2001, 11, 660.
[66] H. Effenberger, Acta Crystallogr. 1992, C48, 2104.
[67] Z. Otwinowski, W. Minor, Methods in Enzymology 1997,
276, 307.
[40] A. Schmidt, Ph. D. thesis, Univ. of Gießen, 2002. URL:
http://bibd.uni-giessen.de/ghtm/2002/uni/d020105.htm.
[68] R. H. Blessing, Acta Crystallogr. 1995, A51, 33.
2578
www.zaac.wiley-vch.de
 2007 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Z. Anorg. Allg. Chem. 2007, 2568Ϫ2578