Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
ARTICLE IN PRESS
N.F. Stephens, P. Lightfoot / Journal of Solid State Chemistry 180 (2007) 260–264
261
which is found to adopt a ‘Super-Diamond’ framework
3. Results and discussion
structure [12].
2. Experimental
2.1. Synthesis
3.1. Framework structure
The [Y3F10]ꢀ framework structure is represented in
Fig. 1(a). This framework is isostructural with that of
(H3O)[Yb3F10] ꢁ H2O12 and may be described, using the
concept of ‘Scale Chemistry’ [15,16] as a ‘Super-Diamond’
network; i.e. a direct analogue of the diamond structure
with each carbon atom being replaced by a [Y6F32]14ꢀ
secondary building unit (SBU). Each yttrium atom is
surrounded by 8 fluorine atoms forming a distorted square
antiprismatic [YF8] primary building unit (PBU), with one
square face markedly smaller than the other one. Six of
these PBUs are linked together via edge sharing, whereby
all 4 edges of the smaller square face are shared with
neighbouring PBUs, forming the octahedral building unit
(SBU) (Fig. 2) at the centre of which is a vacancy. The
[Y6F32] SBUs are further linked into the resultant three-
dimensional structure via the F1 vertices, such that the
resultant framework formula may be represented as
Y6F(2)4F(3)4F(1)24/2ꢃ[Y6F20]2ꢀ, with Z ¼ 8 SBUs per unit
cell. The central vacancy within each SBU is at the position
of a carbon atom in the diamond structure (position 8a (1/
8,1/8,1/8)). For ease of comparison, the ‘classic’ view of the
diamond structure is given in Figure 1(b), and the position
occupied by the [Y6F32] SBU in Fig. 1(c). The final atomic
coordinates obtained from Rietveld refinement of a
completed model are given in Table 1, and bond lengths
in Table 2. The final Rietveld fit is given in Fig. 3.
[C3N2H12]0.5[Y3F10] was prepared by hydrothermal reac-
tion. About 0.0383 g Y(NO3)3 ꢁ 6H2O, 5 mL H2O and 2mL
HF (48% aq.) were heated in a polypropylene bottle at
100 1C for 1 h. The contents were transferred to a Teflon-lined
stainless steel autoclave, washing with 5 mL ethylene glycol.
1.5 mL 1,3-diaminopropane was added to give a gel with a
pH of 6, which was heated at 180 1C for 5 days. The pH at the
end of this period was 6. The white polycrystalline product
was washed with water, filtered and dried at room
temperature. Subsequent attempts to prepare single crystals
suitable for X-ray analysis were unsuccessful. Powder X-ray
diffraction (PXRD) data were collected on a Stoe STADI/P
transmission diffractometer using CuKa1 radiation. Elemen-
tal analysis confirmed phase purity (C 3.73%, H 1.21%, N
2.97% (calc. for C3H12F20N2Y6: C 3.64%, H 1.22%, N
2.83%). Thermogravimetric analysis under N2 revealed a
continuous weight loss of around 2% between room
temperature and 400 1C (perhaps due to a small amount of
surface or intra-crystalline water), followed by an abrupt loss
of about 11% between 400 and 475 1C, corresponding to the
loss of organic moiety plus HF; the residue was identified as
YF3 (PDF no. 32-1431) by PXRD.
2.2. Structure determination
3.2. Template location
In the absence of single crystals, the crystal structure was
determined from the PXRD data. The pattern was indexed
on the basis of the first 20 diffraction maxima, using the
As stated previously, it proved straightforward to locate
the nitrogen atom of the 1,3-diaminopropane cation
directly from the Fourier maps during Rietveld analysis.
However, due to the inherent disorder of the template
moiety in such a high symmetry environment, an un-
ambiguous model of the template position could not be
determined from the PXRD data. The template is located
in a large polyhedral cage defined by 28 fluorine atoms at
the vertices (Fig. 4). The centre of this F28 cage is at a
algorithm of Visser, within the Stoe WinXPOW software
¨
˚
[13]. A face-centred cubic unit cell, with a ꢂ15.48 A was
found, with a figure-of-merit, F20 ¼ 163. A clear resem-
blance to the structure of the lanthanide fluoride structure
type typified by (H3O)[Yb3F10] ꢁ H2O12 was apparent, and
this was subsequently used as a starting model for a
Rietveld analysis, using the GSAS suite [14]. A pseudo-
Voigt peak shape was used, which has five refined
parameters, including a correction for asymmetry at low
2y values. Positions of framework Y and F atoms in space
group Fd3m were input from Ref. [12], and these were
refined, along with the usual lattice, background and
profile parameters. A good-quality Rietveld fit was
obtained at this stage (Rwp ¼ 0.161, w2 ¼ 41). Subsequent
difference Fourier maps clearly revealed (as the highest
peak) a plausible position for the nitrogen atom, within a
large open cage of the structure, and within sensible
hydrogen-bonding distance of framework F atoms. Un-
fortunately, unambiguous location of the remaining C
atoms of the template directly from the PXRD data proved
difficult, due to the inherent disorder expected from the
cubic host symmetry.
3 3 3
position 8b ð8; 8; 8Þ of 43m point symmetry. The faces of the
F28 polyhedron consist of four ‘capped’ hexagons, four
uncapped hexagons and six rectangles. The nitrogen atom,
N(1), located by PXRD lies on a 32e position directly
above the uncapped hexagonal face of the F28 cage, and
forms hydrogen bonds (Table 2) with these six F(1) atoms.
The N(1) position is 50% occupied; in other words, within
any cage, two of the four possible uncapped hexagonal
faces are interacting with the template in this manner. A
plausible template conformation can therefore be proposed
simply by assuming that the central C atom, C(2), of the
1,3-diaminopropane cation lies on the high symmetry 8b
position at the centre of the cage. This leads naturally to
sensible bond lengths within the template moiety by
placing the remaining atom, C(1), at a 96g position shown
Stephens, Nicholas F.
