Exploration on anion ordering, optical properties and electronic structure in K3WO3F3 elpasolite

Article abstract of DOI:10.1016/j.jssc.2011.12.037

Room-temperature modification of potassium oxyfluorotungstate, G2-K ₃WO₃F₃, has been prepared by low-temperature chemical route and single crystal growth. Wide optical transparency range of 0.39.4 μm and forbidden band gap E_g=4.32 eV have been obtained for G2-K₃WO₃F₃ crystal. Meanwhile, its electronic structure has been calculated with the first-principles calculations. The good agreement between the theorectical and experimental results have been achieved. Furthermore, G2-K₃WO₃F₃ is predicted to possess the relatively large nonlinear optical coefficients. 2012 Elsevier Inc.

Full text of DOI:10.1016/j.jssc.2011.12.037

Journal of Solid State Chemistry 187 (2012) 159–164
Contents lists available at SciVerse ScienceDirect
Journal of Solid State Chemistry
journal homepage: www.elsevier.com/locate/jssc
Exploration on anion ordering, optical properties and electronic structure in
K3WO3F3 elpasolite
V.V. Atuchin a, L.I. Isaenko b, V.G. Kesler c, Z.S. Lin d,n, M.S. Molokeev e, A.P. Yelisseyev b, S.A. Zhurkov b
a Laboratory of Optical Materials and Structures, Institute of Semiconductor Physics, SB RAS, Novosibirsk 630090, Russia
b Laboratory of Crystal Growth, Institute of Geology and Mineralogy, SB RAS, Novosibirsk 630090, Russia
c Laboratory of Physical Principles for Integrated Microelectronics, Institute of Semiconductor Physics, Novosibirsk 630090, Russia
d Technical Institute of Physics and Chemistry, Chinese Academy of Sciences, P.O. Box 2711, Beijing 100190, China
e Laboratory of Crystal Physics, Institute of Physics, SB RAS, Krasnoyarsk 660036, Russia
a r t i c l e i n f o
a b s t r a c t
Article history:
Room-temperature modification of potassium oxyfluorotungstate, G2-K3WO3F3, has been prepared by
low-temperature chemical route and single crystal growth. Wide optical transparency range of
Received 17 August 2011
Received in revised form
7 December 2011
0.3–9.4
mm and forbidden band gap Eg¼4.32 eV have been obtained for G2-K3WO3F3 crystal. Mean-
while, its electronic structure has been calculated with the first-principles calculations. The good
agreement between the theorectical and experimental results have been achieved. Furthermore, G2-
K3WO3F3 is predicted to possess the relatively large nonlinear optical coefficients.
& 2012 Elsevier Inc. All rights reserved.
Accepted 26 December 2011
Available online 12 January 2012
Keywords:
Oxyfluorotungstate
Structure
Nonlinear optical crystals
Electronic and optical properties
First-principles calculations
1. Introduction
within the certain temperature intervals [10,13,14,17,18]. Cubic para-
electric phase G0 with elpasolite-type parent structure in space group
Complex metal oxyfluorides are attractive compounds for
creation new noncentrosymmetric crystals because the strong
distortion of metal-(O,F) polyhedra in crystal lattice is achieved
due to different ionicity of metal-O and metal-F bonds. This
strategy is particularly successful in design of new oxyfluorobo-
rates containing boron-(O,F) polyhedra and possessing interesting
nonlinear optical properties and wide transparency range [1–4].
Effects of mixed anion sublattice, however, are considered for a
long time for a suite of transition metals compounds with
[MO6ꢀxFx] (M¼W, Mo, Nb, Ti) groups [5–11]. From geometrical
point of view, the strongest distortion of separated [MO6ꢀxFx]
octahedron is achieved for x¼3 and such polar structural group
seems to be optimal for generation of low-symmetry complex
compounds [6]. These distorted octahedrons can be stacked
without the center of inversion in the crystal lattice if long range
ordering condition is fulfilled. Indeed, several noncentrosymmetric
inorganic crystals containing ordered [MO3F3] (M¼Mo, W) groups
were found at enough low temperatures [6,10,12–16].
Fm-3m has been found for A2BMO3F3 compounds at high tempera-
ture range T4T1 [8,9,16,18–21]. Complete disorder in anion positions
is a principal characteristic of this state. At low temperatures TrT1
partial anion ordering results in subsequent phase transitions with
formation of low-symmetry phases Gi. For many A2BMO3F3 com-
pounds the structural transitions are accompanied by noticeable
dielectric and caloric effects [10,13,14,17,18,20,22–24]. Piezoelectric
Gi-phases were found for several cation compositions but detailed
information on physical properties of the compounds is practically
absent in literature for lack of oxyfluoride single crystals with size and
quality needed for the measurements. To our best knowledge, only
the few-hundred-mm-size Rb2KMoO3F3, K3MoO3F3, (Ag3MoO3F3)(Ag3-
MoO4)Cl, Ag3MoO3F3 and Ag3VO2F4 crystals were grown up to now
because of the great technological difficulty to grow oxyfluoride
crystals with reasonable size [19,25,26].
For the representative compound K3WO3F3 from A2BMO3F3
family, the transitions of ferroelectric G02G1 and ferroelastic
G12G2 character were defined to be at T1¼452 K and T2¼414 K,
respectively [10,14,17,27]. The entropy variation under the tran-
Oxyfluorides A2BMO3F3 (A, B¼Na, K, Rb, Cs, Tl, NH4; M¼Mo, W)
show similar structural, ferroelectric and ferroelastic properties
sition was found to be
D
S1¼RLn(1.68) and
D
S2¼RLn(1.42)
[10,14]. Crystal structures of G1-K3WO3F3 and G2-K3WO3F3
modifications were refined by powder methods in space groups
I4mm and Cm, respectively, with evident indication of O2ꢀ and
Fꢀ ion ordering at anion positions [15,16]. Vibrational spectra and
n Corresponding author. Fax: þ86 10 82543709.
E-mail address: zslin@mail.ipc.ac.cn (Z.S. Lin).
0022-4596/$ - see front matter & 2012 Elsevier Inc. All rights reserved.
doi:10.1016/j.jssc.2011.12.037

160
V.V. Atuchin et al. / Journal of Solid State Chemistry 187 (2012) 159–164
electronic structure of G2-K3WO3F3 were examined with Raman,
IR and X-ray photoelectron spectroscopy (XPS) for powder sam-
ples [27,28]. Noticeable variation of W–O mean bond ionicity in
G2-K3WO3F3, as it seems due to presence of W–F bonds, was
found by XPS in reference to that in pure tungstates [29].
However, other properties of K3WO3F3 remain to be unclear.
Present study is aimed to grow the crystals of low-temperature
ferroelectric G2-K3WO3F3 polymorph. Optical properties and
electronic structure are then evaluated comparatively by experi-
mental and theoretical methods.
2. Experimental
2.1. Chemical synthesis and crystal growth
Powder sample of K3WO3F3 was fabricated by low-tempera-
ture chemical synthesis from water-based solutions. As starting
reagents the KF (99.9%) and K2WO4 (99.5%) were used. At a first
step a solution of potassium fluoride in hydrofluoric acid was
prepared. After this, a solution of potassium tungstate in distilled
water was carried out. Then, these two solutions were mixed with
precautions because the reaction is active and with heating. As a
result of reaction, white precipitate was separated out of mother
solution. After drying a white-colored powder product was
formed.
Fig. 1. Photo of K3WO3F3 single crystal on the background of Times Roman
font 12.
Water-free KF and WO3 were used as starting materials to
produce K3WO3F3 single crystals. Water-free KF was obtained by
evaporating KF ꢁ 2H2O with adding of concentrated etching acid.
NH4F was added to dry the remainder and resulting substance
was heated to 300 1C in a teflon crucible to remove the remained
water and alkali according the following reaction:
KOHþNH4F¼KFþH2OmþNH3m
The obtained substance was heated to 500 1C in closed Pt
crucible in presence of teflon which was used as a fluorine agent.
Then substance was grinded in a mortar, mixed with teflon boring
and placed into a Pt crucible with a cover. The crucible was heated
to 900 1C, the temperature above KF melting temperature, in a
flow of dry nitrogen. Before using, WO3 was heated to 700 1C in a
Pt crucible. The synthesis of K3WO3F3 compound was carried out
from a charge of stoichiometric composition:
Fig. 2. XRD analysis of K3WO3F3 crystal at room temperature. Cm space group.
study for creation of large K3WO3F3 is universal and can be
applied for other oxyfluorides.
3KFþWO3¼K3WO3F3
Phase composition of grown K3WO3F3 crystal at room tem-
perature was evaluated with XRD analysis. The results of the XRD
analysis are shown in Fig. 2. Presence of pure monoclinic
G2-K2WO3F3, space group Cm, modification was found with struc-
The mixture was grinded in a mortar and placed into a Pt
ampule. The ampule was then located on a chamotte support
inside a growth furnace and covered with a Pt cover, under which
dry nitrogen was delivered during the whole growth experiment.
The furnace was heated to 980 1C at a 30 1C/h rate. Then furnace
was kept at this temperature during 6 h and afterwards tempera-
ture was lowered during 2 day at a rate of 1 1C/h. The obtained
crystal was cooled at a rate of 30 1C/h. In a result the cm-size
K3WO3F3 single crystals were formed by this technique as shown
in Fig. 1. The crystals are transparent and enough large for optical
measurements. It should be pointed that presently crystal growth
technology is well developed only for nonlinear optical oxyfluor-
oborates. As to other oxyfluorides, typically powder samples
formed by chemical synthesis are used for physical properties
observation [8,10,13–18,20,23]. For several oxyfluorides tiny
single crystals were fabricated with dimensions enough for single
crystal structure determination [12,19,21,26]. To our best knowl-
edge, up to now preparation of optical quality mm-size crystals
was achieved only for K3MoO3F3, (NH4)2WO2F4 and (NH4)2NbOF5
compounds [25,30,31]. The technology developed in present
˚
˚
˚
tural parameters a¼8.7459(2) A, b¼8.6930(4) A, c¼6.1650(3) A,
¼135.178(2)1.
b
2.2. Structure determination
Structure determination of G0, G1 and G2 polymorph mod-
ifications of K3WO3F3 has been produced with Rietveld analysis
for powder sample prepared by chemical synthesis. The powder
X-ray diffraction patterns were collected at 298, 433 and 513 K
with a Bruker D8 ADVANCE diffractometer in the Bragg-Brentano
geometry and linear Vantec detector [16]. Operating parameters
were: CuK radiation, tube voltage 40 kV, tube current 40 mA,
a
step size 0.0161, counting time 1 s per step. The data were
collected over the range 5–1101. Peak positions were determined
with the program EVA, available in the PC software package
DIFFRAC-PLUS supplied from Bruker. X-ray patterns of the title
compound were indexed using the program McMaille [33]. The
details of the analysis (e.g., R-factors) for the diffraction data sets

V.V. Atuchin et al. / Journal of Solid State Chemistry 187 (2012) 159–164
161
Table 1
Fractional coordinates, occupancies (p) and isotropic atomic displacement parameters (Bi) for K3WO3F3.
2
Atom
p
Wyckoff
positions
Number of atoms
in unit cell
X
Y
Z
˚
Biso.,A
Fm-3, G0
W
1.0
4a
4*1.0¼4
0
0
0
3.90(4)
6.9(1)
5.8(1)
4.3(3)
4.4(7)
K1
1.0
4b
4*1.0¼4
1/2
1/2
1/2
1/4
0
K2
1.0
8c
8*1.0¼8
1/4
1/4
F
0.5
24e
96j
24*0.5¼12
96*0.125¼12
0.219(2)
0.192(3)
0
O
0.125
0.092(2)
0
I4 mm, G1
W
1.0
1.0
1.0
0.5
1
2a
2a
4b
8c
2*1.0¼2
2*1.0¼2
4*1.0¼4
8*0.5¼4
2*1.0¼2
16*0.25¼4
8*0.25¼2
1/2
1/2
1/2
3.82(4)
9.5(3)
5.2(2)
4.1(6)
1.0(4)
4.1(6)
3(1)
K1
0
0
0.497(8)
0.262(2)
0.526(4)
0.217(2)
0.560(4)
0.822(5)
K2
1/2
0
F1
0.277(3)
0
0.277(3)
0
F2
2a
16e
8c
O1
O2
Cm, G2
W
0.25
0.25
0.228(5)
0.110(6)
0.606(5)
0.110(6)
1.0
1.0
1.0
1.0
1.0
0.5
1.0
1.0
0.5
2a
2a
4b
2a
2a
4b
2a
2a
4b
2*1.0¼2
2*1.0¼2
4*1.0¼4
2*1.0¼2
2*1.0¼2
4*0.5¼2
2*1.0¼2
2*1.0¼2
4*0.5¼2
0
0
0
4.5(1)
7.1(9)
1.9(4)
2.0(1)
2.0(1)
2.0(1)
2.0(1)
2.0(1)
2.0(1)
K1
ꢀ0.06(1)
0.006(3)
0.275(7)
0.73(1)
0.04(1)
0.80(1)
0.81(1)
0.04(1)
0.5
ꢀ0.078(8)
0.470(5)
0.07(1)
0.46(1)
0.14(1)
0.60(1)
0.00(1)
0.14(1)
K2
0.719(2)
F1
0
F2
1/2
F3
0.204(7)
O1
O2
O3
0
0
0.204(7)
collected at the three temperatures were reported in Ref. [16], in
which the results have been discussed using the group-theoretical
analysis of the complete order parameter condensate and taking
into account the critical and noncritical atomic displacements. In
present study, however, the structural parameters of known
K3WO3F3 modifications are completed and corrected. Final struc-
tural parameters of G0-, G1- and G2-K3WO3F3 modifications are
shown in Table 1.
2.4. Valence band structure by XPS
Observation of valence band structure of K3WO3F3 was pro-
duced by using surface analysis center SSC (Riber) with X-ray
photoelectron spectroscopy (XPS) method for powder sample
prepared by chemical synthesis. The nonmonochromatic AlK
a
radiation (1486.6 eV) with the power source of 300 W was used
for the excitation of photoemission. The energy resolution of the
instrument was chosen to be 0.7 eV. Under the conditions the
observed full width at half maximum (FWHM) of the Au 4f7/2 line
was 1.31 eV. The binding energy (BE) scale was calibrated in
reference to the Cu 3p3/2 (75.1 eV) and Cu 2p3/2 (932.7 eV) lines,
assuring an accuracy of 0.1 eV in any peak energy position
determination. Photoelectron energy drift due to charging effects
was taken into account in reference to the position of the C 1s
(284.6 eV) line generated by adventitious carbon present on the
surface of the powder as-inserted into the vacuum chamber.
2.3. Optical measurements
Absorption/luminescence spectra for K3WO3F3 were measured
for the plates with thickness of 1.8 mm. The plates were cut and
polished without crystallographic orientation control. Optical
measurements were produced using a UV-2501PC Shimadzu
spectrometer in the UV to near IR region and a Fourier transform
spectrometer Infralum FT-801 in the mid IR. Spectra in the short
wavelength range were measured using a metal cryostat with
silica windows. Absorption coefficients were calculated from
transmission measurements in the transparency region by use
of the following formula [32]:
3. Computational method
In this work, the electronic structures and optical properties
for the G2-K3WO3F3 phase are theoretically studied. The first-
principles calculations were performed using the plane-wave
pseudopotential method [34] implemented in the CASTEP pack-
age [35]. Ultrasoft pseudopotentials [36] are used with the 3s, 3p
and 4s electrons for potassium and the 5s, 5p, 5d and 6s electrons
for tungsten treated as valence electrons. For oxygen and fluorine,
1s electrons are chosen as the core electrons. Generalized-gradi-
ent approximation (GGA) using the Perdew, Burke, and Ernzerhof
(PBE) [37] exchange-correlation functional has been used. The
kinetic energy cutoff is set as 600 eV. Monkhorst–Pack k point
meshes [38] with a density of (5 ꢁ 5 ꢁ 6) points in the Brillouin
zone of the primitive cell are used. Since the position occupancies
for some oxygen and fluorine atom are not unity, the virtual
crystal approximation (VCA) [39] are adopted to take into account
the weight average of the potential of each atomic species at the
position.
2
2
a
ad
T ¼ ð1ꢀRÞ eꢀ d=ð1ꢀR eꢀ2
Þ
ð1Þ
where
a
is the absorption coefficient (cmꢀl), d is the sample
thickness (cm) and R is the power reflection coefficient per
surface. Maximum transmission level allows to estimate roughly
the refraction index n and reflection index R as well as to calculate
. An attempt has been made to determine the dominant
mechanism of band-to-band electronic transitions. The kind of
a
transition can be defined from the functional relationship
between absorption
a and the photon energy hn [32].
Thus, for direct-allowed and indirect-allowed transitions the
relations
a
hn
p(hn
ꢀEg)1/2 and
ahn
p(hvꢀEg)2 are fullfilled,
respectively. The estimations were carried out for a 0.3 mm thick
K3WO3F3 plate. The band gap values Eg were estimated by
2
1/2
extrapolating the linear part of the (
versus h curves to (h
)¼0.
ahn) versus hn or (ahn)
n
a n

162
V.V. Atuchin et al. / Journal of Solid State Chemistry 187 (2012) 159–164
4. Results and discussion
ꢂ825þ375 cmꢀ1, respectively [27]. Maximum transmission in
the range from 4 to 7.5
mm is ꢂ82%, except the band at
The transmission spectra recorded at 300 K for 1.8 mm thick
G2-K3WO3F3 crystal plate is given in Fig. 3(a). The transparency
l
ꢂ5.8 mm. Rough estimation shows that at zero absorption the
reflection coefficients are ꢂ9.5% and 18% for single and multiple
reflections, respectively, corresponding to the refraction indices as
high as nꢂ1.9. We believe that both light scattering and absorp-
range is from 0.3 to 9.4
Transmission grows quickly as wavelength increases but there
is a kink near m after which transmission grows much
ꢂ0.35
slower till m. There is a suite of intensive absorption bands
¼5.0–6.5 m with a maximum at 5.8 m. The
ꢂ0.35–4 m range are due to light scattering. The
lꢀ4 r6
0),
where r is the size of scattering particles (inclusions in our case)
with r5 and e0 are the dielectric permittivity of scattering
mm at the 5% transmission level.
l
m
tion in the K3WO3F3 transparency range from 0.35 to 7.0 mm can
l
ꢂ4
m
be removed and expected transmission spectrum of a high quality
crystal is shown by a dotted line in Fig. 3(a). We established
that fundamental absorption edge rectifies in the coordinates
in the range
l
m
m
losses in the
l
m
intensity of nonresonant Rayleigh scattering is
s
ꢂ
(
e
ꢀ
e
(
a
h
n
)2¼f(h
n) (Fig. 3(b)) and this indicates that direct allowed
electronic transitions determine the absorption edge shape. The
l, e
cross points between the approximation straight line and abscissa
substance and the surrounding medium (crystal), so the scatter-
ing intensity is high at shorter wavelengths and decreases
gradually into the IR spectral region. The intense absorption near
axis (ahn)¼0 give the bandgap value Eg for G2-K3WO3F3 poly-
morph, which is 4.413 and 4.315 eV at 80 and 300 K, respectively.
Such magnitude of Eg shift (ꢂ0.1 eV) with temperature variation
is typical for oxides [42].
l
ꢂ6 mm is likely due to deformation vibrations of the residual
H2O individual molecules which are tied by stronger hydrogen
bonds [40]. These intense bands were also observed in mica, e.g.,
in phlogopite, muscovite: they are associated with the water
molecules adsorbed on the surface and between the structural
layers [41]. In addition, there may be some input of the second
Accounting well known phase transitions in K3WO3F3 at 415
and 452 K [10,14,17,27] one could expect the drastic changes in
the shape and position of the fundamental absorption edge near
these transition temperatures. Nevertheless, our careful measure-
ments on the transmission spectra as a function of temperature
over the range of 300–500 K disagree this anticipation. In Fig. 4
the dependence of spectral position of a point corresponding to a
20% transmission level on temperature is shown. It is clear that
the final curve is smooth enough and there are likely no specific
features or kinks related to phase transitions. Namely, it is
possible to say that even in the presence of phase transitions in
K3WO3F3 the spectrum shape and the bangap values change
gradually. Experiments with heating of K3WO3F3 single crystals
in darkness demonstrated the absence of pyroluminescence
(defined as a spontaneous emission at heating) in the tempera-
ture range from 80 to 600 K. This indicates that no noncentro-
symmetric phases with pyroelectric properties appear/disappear
over the temperature range [43]. However, the presence of
noncentrosymmetric phase without pyroelectric properties, as
in the case of AgGaS2 and AgGaSe2, can not be excluded.
harmonic from intense bands near 825 (12.12
(10.81 m), which are active both in IR absorption and Raman
m
m) and 925 cmꢀ1
m
spectra. Both bands are resulted from the stretching vibrations
of the W–O bond of the [WO3F3]3ꢀ ionic groups [27]. Well-
pronounced bands at
l
ꢂ7.75 (1290 cmꢀ1
)
and 8.38 mm
(1193 cmꢀ1), therefore, may be the sums of the stretching
vibrations of the W–O bond and the stretching vibrations of the
W–F bond in the [WO3F3]3ꢀ ionic groups:ꢂ925þ375 cmꢀ1 and
The G2-K3WO3F3 phase is a polymorph modification with
partial O/F ordering and existence of G3-K3WO3F3 modification
with complete O/F ordering can be reasonably supposed at lower
temperatures To300 K. The stability of Cm phase in K3WO3F3
was investigated by DSM-2M microcalorimeter in the tempera-
ture range of T¼300–110 K. Measurements were carried out in
helium atmosphere on the sample with a mass of 0.2 g in heating
and cooling modes with a rate of 8 K/min. There was not found
Fig. 3. (a) Transmission spectrum of 1.8 mm thick K3WO3F3 single crystal, at
T¼300 K. Dotted curve shows an expected transmission spectrum in inclusions
and water-free K3WO3F3. (b) The fundamental absorption edge for 0.3 mm thick
K3WO3F3 crystal at 80 and 300 K in coordinates (
a
h
n
)2¼f(h
n). Experimental data
are given by points whereas solid lines are a result of approximation in the high
energy range. Cross points between these lines and abscissa axis gives the
bandgap value Eg¼4.413 eV (T¼80 K) and 4.315 eV (300 K).
Fig. 4. A smooth shift of fundamental absorption edge (for points with 20%
transmission) versus temperature. The arrows show position of phase transitions
at 415 and 452 K in K3WO3F3.

V.V. Atuchin et al. / Journal of Solid State Chemistry 187 (2012) 159–164
163
any anomaly behavior of heat capacity which could be associated
with the phase transitions in K3WO3F3 below room temperature.
This probe indicates that phase transition G22G3 may be found
at very low temperatures To110 K if ever occurs.
The calculated electronic structure and the partial density
of states (PDOS) for G2-K3WO3F3 are shown in Fig. 5(a) and (b).
G2-K3WO3F3 is predicted as an indirect gap insulator and the
direct gap at G point is 0.07 eV larger than the indirect band gap.
The calculated energy band gap is 4.2 eV, in good agreement with
the experimental value of 4.32 eV. The PDOS analysis clearly
shows that the O 2p orbitals are located at the valence band
maximum, while the W 5d orbitals are located at the conduction
band minimum. They directly determine the band gap. The
hybridization between the W 5d and O 2p (and F 2p) orbitals
are located in the region from ꢀ4 to ꢀ7 eV. Compared to the
valence band structure measured from XPS (Fig. 6), the good
agreement between the calculated and experimental positions for
the K 3p, F 2s and O 2s states are achieved if the entire calculated
PDOS spectrum is shifted downward by about 5.0 eV. This clearly
confirms the validity of the first-principes method. The further
comparison reveals that the measured peaks appeared at 5.9 eV
and 8.4 eV attribute to the O 2p and W 5d orbitals. It also should
be noted that there are no W 4f orbitals appeared in the
calculated PDOS, since the f orbitals are not included in the
pseudopotential calculations.
Fig. 6. Valence band states in G2-K3WO3F3 as measured with XPS.
Table 2
Nonlinear optical properties of oxyfluoroborates and K3WO3F3 (unit: pm/V).
Composition
dij theo.
Ref.
dij exp.
Ref.
KBe2BO3F2
RbBe2(BO3)F2
CsBe2BO3F2
TlBeBO3F2
BaMgBO3F
BaZnBO3F
0.40
[4]
0.49
0.45
ꢂ0.2
ꢂ0.2
ꢂLBO
50.39
0.23
1.32
ꢂ0.2
0.63
–
[1,45]
[2,45]
[45]
[45]
[46]
[46]
[47]
[3]
Based on the electronic band structures, the refractive indices
and second harmonic generation (SHG) coefficients for G2-
K3WO3F3 are calculated using the formulae in Ref. [44]. At the
Not calculated
Not calculated
Not calculated
0.11
–
–
–
[46]
1.24
[46]
BaCaBO3F
Not calculated
0.74
–
BaAlBO3F2
Ba3Sr4(BO3)3
Ca5(BO3)3F
G2-K3WO3F3
[48]
Not calculated
Not calculated
ꢂ3
–
[49]
[50]
–
This study
This study
radiation wavelength of 1064 nm, we predict that the refractive
indices of G2-K3WO3F3 are nx¼1.776, ny¼1.752, nz¼1.741, and
the birefringence
D
n¼0.035. The calculated refractive indices are
in reasonable relation with estimation nꢂ1.9 from the above
spectroscopic measurements. Moreover, the SHG coefficients are
predicted as d11¼ ꢀ1.97 pm/V, d31¼0.75 pm/V, d12¼3.01 pm/V,
d13¼ ꢀ0.21 pm/V, d32¼0.70 pm/V, and d33¼ ꢀ2.91 pm/V. There-
fore, this crystal has relatively large nonlinear optical coefficients
and may achieve the phase-matching condition, the requirement
of the effective coherent laser output by harmonic generation, in
the mid-IR spectral region. A comparison of nonlinear optical
properties of several known fluoroborates and G2-K3WO3F3
shown in Table 2 reveals that the oxyfluorotungstates could be
the nonlinear optical materials with good performance. A second-
harmonic generation (SHG) test was performed by powder
method under nanosecond pulse pumping at
l¼1.06 mm for the
G2-K3WO3F3 powder sample prepared by low-temperature
chemical synthesis. However, noticeable SHG signal was not
detected supposedly due to phase-matching absence at this
wavelength and mesostructure developed in the powder sample
[15]. The SHG tests on the G2-K3WO3F3 single crystal will be
performed after the sample with bigger size and better optical
quality is obtained. Up to now, the powder SHG signals were
detected only in the G2-Na3MoO3F3 modification where SHG
intensity was about 40 times as large as that of a-quartz [18].
5. Conclusions
The results of present study give new insights into nature and
properties of low-temperature noncentrosymmetric modifications
Fig. 5. Electronic properties for the G2-K3WO3F3 phase. (a) Band structure along
the high symmetry line (b) PDOS projection on the respective orbitals.

164
V.V. Atuchin et al. / Journal of Solid State Chemistry 187 (2012) 159–164
of K3WO3F3 oxyfluorotungstate. Crystal structure is defined for G0,
G1 and G2 phases subsequently formed on cooling from T4452 to
Tꢂ300 K. Complete O/F disorder at anion position is a characteristic
of cubic G0 phase. Partial O/F ordering is obtained for G1 and G2
intermediate phases. Last G3 phase with complete O/F ordering is
not found at T4110 K. Thus, existence of G3-K3WO3F3 modification
is not confirmed up to now.
Effective technology developed for crystal growth of low-
temperature G2 modification yields crystals suitable for optical
measurements. As a result optical transparency range of 0.3–9.4 mm
and band gap Eg¼4.32 eV were first time defined for G2-K3WO3F3
crystal. This indicates that oxyfluorides are transparent in UV
spectral range and this chemical class is promising for creation of
new UV materials.
DFT calculations reveals that in G2-K3WO3F3 the indirect band
gap is governed by the O 2p orbitals located at the valence band
maximum and the W 5d orbitals located at the conduction band
minimum. Reasonable level of nonlinear optical coefficients and
comparatively low birefringence are predicted for G2-K3WO3F3,
space group Cm, by model calculations. The results, however, show
a possibility of nonlinear optical properties for oxyfluorotungstates.
[7] K.R. Heier, et al., Inorg. Chem. 38 (1999) 762.
[8] M.S. Molokeev, et al., Powder Diffr. 22 (2007) 227.
[9] M.R. Marvel, et al., J. Am. Chem. Soc. 129 (2007) 13963.
[10] V.D. Fokina, et al., Ferroelectrics 347 (2007) 60.
[11] M.R. Marvel, et al., Z. Anorg. Allg. Chem. 635 (2009) 869.
[12] P.A. Maggard, et al., J. Solid State Chem. 175 (2003) 27.
[13] G. Pe´raudeau, et al., Solid State Commun. 27 (1978) 515.
[14] J. Ravez, et al., Ferroelectrics 26 (1980) 767.
[15] V.V. Atuchin, et al., Chem. Phys. Lett. 493 (2010) 83.
[16] M.S. Molokeev, et al., Phys. Solid State 53 (2011) 834.
[17] G. Peraudeau, et al., Solid State Commun. 27 (1978) 591.
[18] J.P. Chaminade, et al., Mater. Res. Bull. 21 (1986) 1209.
[19] S.C. Abrahams, et al., Acta Crystallogr.. B 37 (1981) 1332.
[20] F.J. Brink, et al., J. Solid State Chem. 174 (2003) 44.
[21] A.A. Udovenko, et al., Acta Crystallogr. B 64 (2008) 305.
[22] I.N. Flerov, et al., Ferroelectrics 346 (2007) 77.
[23] I.N. Flerov, et al., Phys. Solid State 50 (2008) 515.
[24] I.N. Flerov, et al., Crystallogr. Rep. 56 (2011) 9.
[25] Z.G. Ye, et al., Ferroelectrics 124 (1991) 281.
[26] J.M. Chamberlain, et al., Cryst. Growth Des. 10 (2010) 4868.
[27] A.A. Ekimov, et al., Ferroelectrics 401 (2010) 168.
[28] V.V. Atuchin, et al., Solid State Commun. 150 (2010) 2085.
[29] V.V. Atuchin, et al., Cryst. Growth Des. 11 (2011) 2479.
[30] S.V. Mel’nikova, et al., Phys. Solid State 48 (2006) 117.
[31] S.V. Mel’nikova, et al., Phys. Solid State 52 (2010) 2168.
[32] T.S. Moss, Optical Properties of Semiconductors, Butterworth, London, 1961.
[33] A. Le Bail, Powder Diffr. 19 (2004) 249.
[34] M.C. Payne, et al., Rev. Mod. Phys. 64 (1992) 1045.
[35] S.J. Clark, et al., Z. Kristallogr. 220 (2005) 567.
[36] D. Vanderbilt, Phys. Rev. B 41 (1990) 7892.
Acknowledgments
[37] J.P. Perdew, et al., Phys. Rev. Lett. 77 (1996) 3865.
[38] H.J. Monkhorst, et al., Phys. Rev. B 13 (1976) 5188.
[39] L. Bellaiche, et al., Phys. Rev. B 61 (2000) 7877.
[40] V.F. Kokorina, Glasses for Infrared Optics, CRC Press, Boca Raton, FL, USA,
1996., p. 236.
[41] T.I. Shishelova, et al., Workshop on Spectroscopy. Water in minerals,
Akademiya estestvoznaniy, 2010, (in Russian) ISBN 978-5-91327-093-1.
[42] J. Aarik, et al., Thin Solid films 466 (2004) 41.
This study is partly supported by SB RAS Project No. 34. ZSL
acknowledges the funding support of No. 91022036, 11174297,
2010CB630701, and 2011CB922204 in China. Authors also wish to
thank V. D. Fokina for microcaloric measurements.
[43] A.P. Yelisseyev, et al., Avtometriya 4 (1988) 112. in Russian.
[44] C.T. Chen, et al., Appl. Phys. B 80 (2005) 1.
[45] C.D. McMillen, et al., J. Cryst. Growth 310 (2008) 2033.
[46] R.K. Li, et al., Inorg. Chem. 49 (2010) 1561.
References
[1] C.T. Chen, et al., Appl. Phys. B 97 (2009) 9.
[2] C.T. Chen, et al., J. Opt. Soc. Am. B 26 (2009) 1519.
[3] Y. Zhou, et al., Opt. Express 17 (2009) 20033.
[4] Z.S. Lin, et al., J. Appl. Phys. 109 (2011) 073721.
[5] G. Pausewang, et al., Z. Anorg. Allg. Chem. 364 (1969) 69.
[6] J. Ravez, J. Phys. III Fr. 7 (1997) 1129.
[47] X. Wang, et al., Opt. Mater. 29 (2007) 1658.
[48] H. Huang, et al., J. Appl. Phys. 106 (2009) 103107.
[49] G.C. Zhang, et al., Cryst. Growth Des. 9 (2009) 3137.
[50] K. Xu, et al., Opt. Express 16 (2008) 17735.