Unprecedented zeolite-like framework topology constructed from cages with 3-rings in a barium oxonitridophosphate

Article abstract of DOI:10.1021/ja202159e

A novel oxonitridophosphate, Ba₁₉P₃₆O _6+xN_66-xCl_8+x (x ≈ 4.54), has been synthesized by heating a multicomponent reactant mixture consisting of phosphoryl triamide OP(NH₂)₃, thiophosphoryl triamide SP(NH₂)₃, BaS, and NH₄Cl enclosed in an evacuated and sealed silica glass ampule up to 750 °C. Despite the presence of side phases, the crystal structure was elucidated ab initio from high-resolution synchrotron powder diffraction data (λ = 39.998 pm) applying the charge flipping algorithm supported by independent symmetry information derived from electron diffraction (ED) and scanning transmission electron microscopy (STEM). The compound crystallizes in the cubic space group Fm3c (no. 226) with a = 2685.41(3) pm and Z = 8. As confirmed by Rietveld refinement, the structure comprises all-side vertex sharing P(O,N)₄ tetrahedra forming slightly distorted 3⁸4⁶8¹² cages representing a novel composite building unit (CBU). Interlinked through their 4-rings and additional 3-rings, the cages build up a 3D network with a framework density FD = 14.87 T/1000 A³ and a 3D 8-ring channel system. Ba²⁺ and Cl^- as extra-framework ions are located within the cages and channels of the framework. The structural model is corroborated by ³¹P double-quantum (DQ) /single-quantum (SQ) and triple-quantum (TQ) /single-quantum (SQ) 2D correlation MAS NMR spectroscopy. According to ³¹P{¹H} C-REDOR NMR measurements, the H content is less than one H atom per unit cell.

Full text of DOI:10.1021/ja202159e

ARTICLE
pubs.acs.org/JACS
Unprecedented Zeolite-Like Framework Topology Constructed from
Cages with 3-Rings in a Barium Oxonitridophosphate
Stefan J. Sedlmaier, Markus D€oblinger, Oliver Oeckler, Johannes Weber, J€orn Schmedt auf der G€unne,
and Wolfgang Schnick*
Ludwig-Maximilians-Universit€at M€unchen, Department Chemie, Butenandtstrasse 5-13 (D), D-81377 Munich, Germany
S Supporting Information
b
ABSTRACT: A novel oxonitridophosphate, Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54), has been
synthesized by heating a multicomponent reactant mixture consisting of phosphoryl triamide
OP(NH2)3, thiophosphoryl triamide SP(NH2)3, BaS, and NH4Cl enclosed in an evacuated
and sealed silica glass ampule up to 750 ꢀC. Despite the presence of side phases, the crystal
structure was elucidated ab initio from high-resolution synchrotron powder diffraction data
(λ = 39.998 pm) applying the charge flipping algorithm supported by independent symmetry
information derived from electron diffraction (ED) and scanning transmission electron
microscopy (STEM). The compound crystallizes in the cubic space group Fm3c (no. 226)
with a = 2685.41(3) pm and Z = 8. As confirmed by Rietveld refinement, the structure
comprises all-side vertex sharing P(O,N)4 tetrahedra forming slightly distorted 3846812 cages
representing a novel composite building unit (CBU). Interlinked through their 4-rings and
additional 3-rings, the cages build up a 3D network with a framework density FD = 14.87 T/
1000 Å3 and a 3D 8-ring channel system. Ba2+ and ClÀ as extra-framework ions are located
within the cages and channels of the framework. The structural model is corroborated by 31P double-quantum (DQ) /single-quantum
(SQ) and triple-quantum (TQ) /single-quantum (SQ) 2D correlation MAS NMR spectroscopy. According to 31P{1H} C-REDOR
NMR measurements, the H content is less than one H atom per unit cell.
’ INTRODUCTION
versions (e.g., MeAPO, MeAPSO) thereof.2,5 Thus, the field of
zeolite chemistry seems quite mature which means that the search
for new framework types becomes increasingly challenging.
The exchange of oxygen by nitrogen in the anionic substruc-
ture is an innovative but rarely realized expansion of zeolite
chemistry. Nitrido-zeolites promise beneficial chemical and
physical properties (e.g., higher thermal stability or adjustable
acidity/basicity) and a huge structural diversity. As compared
with oxygen, nitrogen atoms are more common in three-binding
situations, and they provide more flexibility as bridging atoms in
networks by occasionally realizing smaller angles TÀXÀT (X =
O, N). Consequently, both large rings as well as rare 3-rings can
be stabilized so that novel zeolite-like frameworks become
possible.
Classical zeolites, such as aluminosilicates and aluminophos-
phates, are well-established in fundamental industrial processes,
e.g., substance separation, air and water conditioning, or catalysis.
As they have the potential for further applications in future
technologies (e.g., sensors, electronic, or optical systems), inor-
ganic open-framework materials emerged as a research area with
a multitude of compound classes in the last decades. In addition
to diverse metal phosphates, germanates, and borates, there are
sulfates, arsenates, or phosphonates as well as organicÀinorganic
hybrid compounds with porous networks.1,2 However, many
microporous structures are thermally and chemically not suffi-
ciently stable to make their way toward advanced materials.
Consequently, it is worthwhile to synthesize novel open-frame-
work materials that exhibit three-dimensional, rigid framework
structures based on vertex-sharing tetrahedra. Since the discov-
ery of the aluminophosphates by Flanigen et al.3 in the 1980s, it
has been attempted with great creativity and effort to access new
stable frameworks with different pore sizes and shapes combined
with varying chemical and physical properties. Different synthe-
sis conditions (temperature, reaction time, pH), many different
structure-directing agents (SDA), as well as a broad spectrum of
solvents, including ionic liquids, were employed. The fluoride
route4 has been utilized, and other tetrahedra centers (e.g., B, Ga,
Zn) were included, resulting in new zeotypes in compound classes
like silicoaluminophosphates (SAPOs) and metal-containing
This nitride concept became reality in (oxo-)nitridosilicates
and (oxo-)nitridophosphates. After the proof of concept with
nitridosodalites6 and related oxonitridosodalites,7 the benefits of
nitrogen in zeolite-like framework structures have been demon-
strated only for very few examples. Besides a zeolite-like SiÀN
framework in Ba2Nd7Si11N23 with a notable thermal stability up to
1600 ꢀC,8 the flexibility of N bridging resulted in LixH12ÀxÀy+z
-
[P12OyN24Ày]Xz for X = Cl, Br with a new zeolite topology,
namely, NPO (nitridophosphate one).9 The typical (ring-)strain
Received: March 20, 2011
Published: June 24, 2011
r
2011 American Chemical Society
12069
dx.doi.org/10.1021/ja202159e J. Am. Chem. Soc. 2011, 133, 12069–12078
|

Journal of the American Chemical Society
ARTICLE
in such an extended 3-ring system could only be realized with that
framework type material and the related nitridic compounds
Ba3Si3N5OCl, Ba3Ta3N6Cl, Ba15Ta15N33Cl4, and Ba6Si6N10O2-
(CN2) so far.10,11 The clathrate P4N4(NH)4(NH3)12 with its
4284 cages encapsulating ammonia molecules could be synthe-
sized because rather small angles TÀXÀT can be realized in
phosphorus nitride networks. Taking into account the high
thermal and chemical stability and further proposed applications
like, e.g., the clathrate as gas storage or membrane reactor
material,13 these examples show the potential of the nitride
chemistry in the field of open-framework structures.
In this contribution, we describe the synthesis and structure
elucidation of the oxonitridophosphate Ba19P36O6+xN66ÀxCl8+x
(x ≈ 4.54) which exhibits a novel zeolite-like framework
topology involving a new composite building unit (CBU).
’ EXPERIMENTAL SECTION
Figure 1. Observed (crosses) and calculated (gray line) powder
diffraction pattern as well as difference profile of the Rietveld refinement.
Peak positions are marked by vertical lines. Regions with impurity
reflections have been excluded (wherever applicable, see text).
Synthesis of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54). Phosphoryl
triamide OP(NH2)3 and thiophosphoryl triamide SP(NH2)3 were
synthesized according to the literature14,15 similar to procedures de-
scribed by Correll.16 First, 10À20 mL of freshly distilled POCl3 (Acros
Organics, Geel, Belgium, 99%) and PSCl3 (Acros Organics, 98%),
respectively, were added directly and slowly to liquid ammonia in a
flame-dried Schlenk-type 1 L flask. In the second step, the elimination of
NH4Cl from the products was carried out performing a Soxhlet
extraction with distilled Et2NH (Gr€ussing GmbH, Filsum, Germany,
99%) in dry CH2Cl2 for three days. After drying in vacuo, OP(NH2)3
and SP(NH2)3 are available as starting materials in the form of colorless,
water-sensitive powders. Their purity was verified with powder X-ray
diffraction.
unequivocally. Indexing by the SVD method22 yielded cubic unit cells,
either cP with a = 1341.62 pm or—taking into account some additional
weak reflections—cF with a = 2683.24 pm.
For the further process, the pattern with the lowest amount of
impurity reflections was selected. After evaluation of scanning transmis-
sion electron microscopy (STEM) images recorded in high-angle
annular dark field (HAAF) mode and selected-area electron diffraction
(SAED) patterns (see below), the space group Fm3c (no. 226) was
unambiguously identified. The structure solution succeeded ab initio
using the charge flipping algorithm23 and subsequent difference Fourier
syntheses and placing bridging atoms at reasonable positions. Rietveld
refinement of the final structure model was carried out applying the
fundamental parameters approach (direct convolution of source emis-
sion profiles, instrument contributions, and crystallite size and micro-
strain effects).24 The preferred orientation of the crystallites was
described with fourth-order spherical harmonics. As far as possible,
regions containing parasitic reflections from byproducts have been
excluded. Remaining misfits indicated by the difference profile originate
from regions where impurity reflections are overlapping with those of
the target compound (cf. Figure 1). Overall displacement parameters
have been used for ClÀ and N/O, respectively. A common O/N ratio on
all bridging atom positions was derived by constraining the occupancy to
be 100% and guaranteeing charge neutrality by taking the partially
occupied ClÀ positions into account. For an optimized tetrahedral
geometry, distance constraints (165 pm for PÀ(N/O), 269 pm for
(N/O)À(N/O)) have been included. Furthermore, absorption as well
as a cylindrical 2θ correction25 was applied. Crystallographic data and
further details of the data collection are summarized in Table 1. Table 2
shows the positional and displacement parameters for all atoms. Due to a
step width of 0.002ꢀ2θ, severe serial correlation occurs. Consequently,
the estimated standard derivations are underestimated.26,27 The Rietveld
fit is displayed in Figure 1. Further details of the crystal structure
investigation may be obtained from Fachinformationszentrum Karls-
ruhe, 76344 Eggenstein-Leopoldshafen, Germany (fax: (+49) 7247-
808-666; e-mail: crysdata(@)fiz-karlsruhe.de, http://www.fiz-karlsruhe.
de/request_for_deposited_data.html) on quoting the depository num-
ber CSD-422769.
Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) has been synthesized employing
a multicomponent reactant system. In a typical procedure, BaS (57.9 mg,
0.342 mmol; Sigma-Aldrich, 99.9%), OP(NH2)3 (10.3 mg, 0.108 mmol),
SP(NH2)3 (60.0 mg, 0.540 mmol), and NH4Cl (23.1 mg, 0.432 mmol;
Fluka, puriss. p.a.) were thoroughly mixed and ground in a glovebox
(MBraun, Garching, Germany) and subsequently transferred into a silica
glass ampule (wall thickness 2 mm, inner diameter 11 mm). The
evacuated and sealed ampule (length around 110 mm) was heated to
200 and 750 ꢀC in a tube furnace with dwell times of 12 and 48 h
(heating rate, 1 K minÀ1; cooling rate, 0.25 K minÀ1), respectively. The
emerging condensation products NH3 and H2S partially are deposited as
(NH4)2S together with excess NH4Cl at the cooler zones of the ampule.
After breaking the ampules, the samples were washed with water and
DMF to remove the remaining NH4Cl and BaS. Ba19P36O6+xN66ÀxCl8+x
(x ≈ 4.54) was obtained as a colorless, water- and air-resistant,
microcrystalline powder containing small amounts of various crystalline
and amorphous byproducts that could not be removed by washing.
Besides the starting materials named above, BaCl2 or other P/N/O/Cl
containing chemicals like P(NH2)4Cl,17 [PN(NH)2]318 or (NH2)2(O)-
19
PNP(NH2)3 can also be used in the synthesis. The purest product,
however, was obtained with the combination named at the beginning of
the paragraph.
Powder X-ray Diffraction (PXRD), Structure Solution, and
Rietveld Refinement. High-resolution synchrotron PXRD data of
different samples were collected at 298 K at beamline ID31 (ESRF,
Grenoble, France), using a DebyeÀScherrer setup (with spinning glass
capillaries, 1 mm diameter) with monochromatic radiation (λ = 39.998 pm)
and a nine-crystal multianalyzer detector.20
Extraction of the peak positions, pattern indexing, structure solution,
Fourier calculations, and Rietveld refinements was carried out with the
TOPAS package.21 By evaluation of all measured patterns, the reflec-
tions of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) have been identified
Electron Microscopy. SAED and STEM measurements were
carried out on a transmission electron microscope FEI Titan 80-300
equipped with a field emission gun operating at 300 kV. STEM-HAADF
12070
dx.doi.org/10.1021/ja202159e |J. Am. Chem. Soc. 2011, 133, 12069–12078

Journal of the American Chemical Society
ARTICLE
Table 1. Crystallographic Data for Ba19P36O6+xN66ÀxCl8+x
outer diameters of 2.5 or 1.3 mm which were mounted in commercial
MAS probes (Bruker). The rotors were spun at rotation frequencies
νMAS between 8 and 20 (2.5 mm) and 50 kHz (1.3 mm), respectively.
Chemical shifts are given relative to the respective reference compounds
(1H: 1% tetramethylsilane (TMS) in CDCl3; 31P: 85% H3PO4, T =
298 K) as external standards. The calibration of the spectrometer was
done with TMS under MAS conditions using the unified scale Ξ and the
chemical shift definitions in the literature.29 Quantitative 1H spectra and
31P spectra were obtained by a rotor-synchronized spinÀecho sequence
at νMAS = 20 kHz with a repetition delay of 32 s in the case of 1H and of
1500 s in the case of 31P.
(x ≈ 4.54) in Fm3c (esd’s in Parentheses)
crystal structure data
formula
formula mass/g molÀ1
Ba19P36O10.54N61.46Cl12.54
5198.4143
crystal system
cubic
space group
Fm3c (no. 226)
a = 2685.41(3)
V = 19365.6(6)
Z = 8
cell parameter/pm
cell volume/106 pm3
formula units/cell
A
31P 2D double-quantum (DQ) single-quantum (SQ) correlation
MAS NMR spesctrum was obtained using PostC730 with a conversion
period of 0.8 ms and rotor-synchronized data sampling of the indirect
dimension. A 31P 2D triple-quantum (TQ) single-quantum (SQ)
correlation MAS NMR spectrum31 was obtained with a conversion
period of 1.6 ms and rotor-synchronized data sampling of the indirect
dimension. A 31P{1H} PostC313-REDOR32 curve was recorded to
determine the effective dipolar coupling constant and estimate limits
for the interatomic PÀH distances. The experimental data are compared
with simulated curves for different effective dipolar couplings on the
universal dephasing scale.33
X-ray density/g cmÀ3
FD/T10À3 ÅÀ3
F = 3.566(2)
14.87
3
data collection
radiation
synchrotron (beamline ID31,
ESRF, Grenoble), λ = 39.998 pm
temperature/K
298(2)
1.0À45.5
22000
2θ range/ꢀ
data points
number of observed reflections
3211
structure refinement
’ RESULTS AND DISCUSSION
method of refinement
program used
fundamental parameters model24
TOPAS-Academic 4.121
Synthesis. Oxonitridophosphates are isolobal and isoelec-
tronic with silicates. However, the variety of silicate structures
is far from being emulated with P/O/N chemistry. This suggests
great potential, but the reaction and crystallization temperatures
are often close to the decomposition temperature of the starting
materials (e.g., PON and P3N5 decompose from 800 ꢀC) or the
(oxo-)nitridophosphate phases themselves.34 To prevent this
decomposition,35 high-pressure synthesis36 has been used for
most known P/(O)/N phases. However, high pressure obviously
favors materials with a high density and is not the method
of choice to obtain porous open-framework oxonitridophos-
phates. Therefore, suitable starting materials and reaction con-
ditions are required, e.g., P/(O)/N/H-containing molecular
compounds in the closed system of evacuated silica glass ampules
in the presence of metal salts and NH4X (X = halogen) as
mineralizer.6,7,9
atomic parameters
profile parameters
background function/parameters
other parameters
27
9
shifted Chebyshev/40
4
restraints
3
constraints
15
fwhm (reflection at 3.424ꢀ2θ)/ꢀ
0.008
χ2 = 5.172
Rp = 0.0906
wRp = 0.1215
RBragg = 0.0770
R indices
images were recorded using a Fischione model 3000 ADF detector, and
diffraction patterns were recorded with a Gatan UltraScan 1000 (2k Â
2k) CCD camera. The samples were finely dispersed in ethanol by
sonication, and a small amount of the suspension was subsequently
dispersed on copper grids coated with holey carbon film. The grids were
mounted on a double tilt holder with a maximum tilt angle of 30ꢀ.
Simulations of electron diffraction patterns were carried out with the
online version of the EMS program package (Electron Microscopy
Image Simulation).28
Scanning electron microscopy (SEM) and energy dispersive X-ray
analysis (EDX) were performed using a JSM-6500F electron microscope
(JEOL Ltd., Tokyo, Japan) with a field emission source equipped with an
EDX detector model 7418 (Oxford Instruments, Oxfordshire, UK).
Powders were placed on a brass sample carrier fixed with self-adhesive
carbon plates (Plano GmbH, Wetzlar, Germany). The samples were
sputtered with carbon (sputter device: BAL-TEC MED 020, BAL-TEC
AG, Balzers, Netherlands). EDX data collection and evaluation was
carried out with the aid of the INCA program package.
Besides other starting materials like (NH2)2(O)PNP(NH2)3,19
a convenient approach for synthesis of oxonitridophosphate
frameworks is the joint thermal condensation of phosphoryl
triamide OP(NH2)3 and thiophosphoryl triamide SP(NH2)3.
While the latter allows a controlled addition of oxygen, the
ammonia partial pressure emerging during the condensation in a
closed system offers the conditions necessary for the crystal-
lization process which can further be adjusted by adding a
respective amount of NH4X (X = halogen). The metal salts act
as structure-stabilizing agents.
In the case of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54), the
combination of BaS, OP(NH2)3, SP(NH2)3, and NH4Cl
(cf. Experimental Section) led to the best result after a series of
experiments with different starting materials (e.g., BaCl2 or
P(NH2)4Cl) in different molar ratios, and different temperature
programs had been carried out to optimize the yield of the title
compound. Although Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) is
routinely obtained by different approaches, it turned out that
the reactant system is very sensitive concerning the appearance
and type of impurities. This is an intrinsic problem as at least four
starting materials react at high temperatures. Many parameters
Solid-State MAS (Magic Angle Spinning) NMR (Nuclear
Magnetic Resonance) Methods. 1H and 31P solid-state MAS NMR
spectra of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) were recorded at ambient
temperature on a Bruker Avance III spectrometer with an 11.75 T
magnet. The powdered samples were contained in ZrO2 rotors with
12071
dx.doi.org/10.1021/ja202159e |J. Am. Chem. Soc. 2011, 133, 12069–12078

Journal of the American Chemical Society
ARTICLE
Table 2. Atomic Coordinates, Wyckoff Symbols, and Isotropic Displacement Parameters Biso/Å2 for the Atoms in Ba19P36O6+x
66ÀxCl8+x (x ≈ 4.54) in Fm3c (esd’s in Parentheses Are Underestimated Due to Severe Serial Correlation)
-
N
atom
Wyckoff symbol
x
y
z
occupancy
Biso
Ba(1)
Ba(2)
Ba(3)
P(1)
48f
96i
8b
0.1055(1)
0
1/4
1/4
1
2.28(5)
3.06(5)
1.56(13)
0.58(8)
1.81(16)
0.1402(1)
0.1334(1)
1
0
0
0
1
192j
96h
64g
96i
0.6238(2)
1/4
0.6182(1)
0.6962(1)
1
P(2)
0.9441(1)
y
1
Cl(1)
Cl(2)
N(1)
O(1)
N(2)
O(2)
N(3)
O(3)
N(4)
O(4)
0.0689(2)
0.2969(5)
x
x
1
1.14(11)
0
0.7593(7)
0.378(5)
0.854(5)
0.146(5)
0.854(5)
0.146(5)
0.854(5)
0.146(5)
0.854(5)
0.146(5)
96i
0
0.9331(2)
0.3291(2)
0.8591(2)
0.5712(2)
0.7337(2)
0.4025(2)
y
192j
96h
192j
0.3334(2)
1/4
2.56(20)
0.5840(2)
0.6984(2)
Indexing and Structure Solution. As no single crystals of
adequate dimensions were available, it was a challenge to solve the
structure from PXRD data obtained from multiphase samples of
Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54). Starting with data from
conventional laboratory diffractometers, indexing was not possible
due to low resolution (with Mo KR radiation) or a bad signal-to-
noise ratio (with Cu KR radiation). An unambiguous determina-
tion of the unit cell was additionally impeded by several reflections
with low intensity at low angles; it was unclear whether the latter
are superstructure reflections or due to unknown side phases.
Indexing became possible and fairly unambiguous after evaluating
of high-resolution synchrotron PXRD data from different samples.
As each sample contains different impurity phases, the reflections
of the main phase Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) could be
clearly identified. While initially a cP unit cell with a = 1341.62 pm
had been considered, a cF unit cell with a doubled latticeparameter
was suggested by reflections that could be superstructure reflec-
tions (h, k, l = 2n + 1). Although the doubling of a results in an
8-fold cell volume due to the F-centering, this, however, exhibits a
2-fold superstructure. To exclude the possibility that these reflec-
tions are due to unknown impurities, the 2-fold superstructure was
subsequently confirmed by SAED. Figure 3 compares the experi-
mental diffraction pattern of the [0À15] zone axis with respective
simulations on the basis of structure models in the primitive (a =
1342.95 pm) and in the F-centered (a = 2685.41 pm) cell. The
appearances of reflectionson odd lines, e.g., reflectionswithindices
h15, requirethe doubled latticeparameter. A further comparison of
a simulated and an experimental electron diffraction pattern is
indicated in the Supporting Information (Figure S1).
Figure 2. SEM images of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) as
cuboidal microcrystals and the Ba/P/O/N side phase as nanometer-
sized platelets (bright areas in the big image and small SEM image).
As some cF space groups are indistinguishable by means of
reflection conditions, STEM-HAADF images along [100] and
[110] (Figure 4, left) have been used to derive the correct space
group. Comparison of their plane groups (p4mm (top) and
p2mm (bottom)) with the symmetry of the special projections
leaves only two options (Fm3c, Fd3c). As Fd3c is not consistent
with the extinction conditions analyzed in the PXRD pattern,
Fm3c is unambiguously the correct space group.
such as the NH3 partial pressure and the amount of ClÀ ions are
simultaneously affected and cannot be fully controlled, so far
impeding the phase-pure synthesis of Ba19P36O6+xN66ÀxCl8+x
(x ≈ 4.54). However, the title compound is typically obtained as
the most prominent crystalline phase with more than 90% yield
(estimated from the intensity ratio of strongest Bragg peaks of the
title compound and side phase(s)). Figure 2 shows the title
compound as cuboidal microcrystals with a mean edge length of
3À5 μm together with the most significantside phase, an unknown
Ba/P/O/N compound in the form of nanometer-sized platelets.
Structure solution succeeded ab initio by applying the charge-
flipping algorithm.23 Initially, the Ba, P, and fully occupied Cl
positions have been located. The correctness of the heavy-atom
12072
dx.doi.org/10.1021/ja202159e |J. Am. Chem. Soc. 2011, 133, 12069–12078

Journal of the American Chemical Society
ARTICLE
Figure 3. (a) SAED pattern of a cuboidal microcrystal of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) (top) and corresponding kinematic simulations of
electron diffraction patterns in the [0À15] zone axis for structure models refined in space groups Pm3m (b) and Fm3c (c).
directly identified because of the atomic number sensitivity of the
STEM-HAADF in [100] and [110]. Figure 4 depicts the match
of the heavy-atom structure with the STEM images. The final
crystal structure has been elucidated by evaluation of difference
Fourier maps generated after refinements of fragmentary struc-
ture models and placing additional bridging atoms at reasonable
positions. The final Rietveld refinement of the complete struc-
ture model was performed in space group Fm3c. This is reason-
able as almost all superstructure reflections are observed,
although they are rather weak. Additionally, refinement of the
basic structure in space group Pm3m was done as well. The results
are given in the Supporting Information (Tables S1 and S2).
NMR Study. In addition to the X-ray/electron diffraction data,
solid-state NMR spectroscopic investigations have been per-
formed. These experiments can add further information about
the structure, for example, if hydrogen is present in the title
compound. Thereby, additional evidence for the structural
model of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) has been obtained
by an independent method.
A quantitative 31P NMR spectrum is shown at the top of the
Figure 4. STEM-HAADF images (left) indicating the heavy atom structure
(right) of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54). Top: image along [100] with
plane group p4mm. Bottom: image along [110] with plane group p2mm.
2D spectrum in Figure 5. Two peaks are observed at À4.6
and À17.7 ppm. These shift values agree well with NMR data of
most other nitridophosphates36 and oxonitridophosphates
(oxonitridosodalite, À8.7 ppm;7 NPO type material, 6.5, À0.1
ppm;9 Sr3P6O6N8, 2.6 ppm).37 With the resolution power of a
31P double-quantum filtered 2D spectrum, several independent
positions was corroborated by the STEM-HAADF images.
Incorrect solutions regarding heavy atom positions could be
12073
dx.doi.org/10.1021/ja202159e |J. Am. Chem. Soc. 2011, 133, 12069–12078

Journal of the American Chemical Society
ARTICLE
Figure 6. 31P-MAS NMR triple-quantum (TQ) single-quantum (SQ)
correlation spectrum of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54). The 1D
spectrum (top) was obtained by direct excitation; the diagonal is
indicating the position of TQ coherences of three isochronous nuclei;
and the peaks labeled with AÀAÀA, AÀAÀB, AÀBÀB, and BÀBÀB
belong to the title compound. The amorphous impurity is almost
perfectly suppressed.
Figure 5. 31P-MAS NMR double-quantum (DQ) single-quantum
(SQ) correlation spectrum of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54). The
quantitative 1D spectrum (top) was obtained by direct excitation, and
below are the individual line shapes of the four Gaussian/Lorentzian profiles
which were obtained by constrained line shape deconvolution (dashed
lines, amorphous; dotted lines, crystalline). The diagonal is indicating the
position of DQ coherences of two isochronous nuclei. The sharp DQ
coherences of two crystalline 31P sites are visible next to a broad
autocorrelated DQ peak which is assigned to an amorphous impurity.
results, the line shape parameters were first constrained on the
sum projection of the AÀB DQ resonance in the 2D spectrum.
Subsequently, the quantitative spinÀecho spectrum was decon-
voluted which resulted in a 5.2:10.0 peak area ratio for peaks A
and B, respectively. The deconvolution data in Table 3 also show
that—besides P-containing impurities—a high proportion of the
sample is amorphous or heavily disordered. In agreement with
the SEM image (Figure 2), we conclude that during the reaction
first amorphous particles are formed and later crystallization sets
in from the particle surface.
Table 3. Peak Assignment in the Quantitative SpinÀEcho 31
NMR Spectrum of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54)
P
δiso/ppm
peak area/au
T1/s
assignment
To clarify whether the sharp and broad peaks originate from
the same phase, for example, in the form of a crystal with regions
of heavy and weak N/O disorder, a 31P triple-quantum filtered
2D spectrum (Figure 6) was acquired. This cannot be concluded
from the 31P double-quantum filtered 2D spectrum because of
the spectral overlap of peak A with the broad features. 31P triple-
quantum coherences, which are expected for the crystalline title
compound, are only observed between peaks A and B, while the
broad contributions centered around À3 and À9 ppm are
reduced in intensity. It is noteworthy that the triple-quantum
coherences involving phosphorus atoms P(1) and P(2) are not
related to the broad features. Hence, the description of the
sample as a heterogeneous mixture of the title compound and an
amorphous impurity is consistent with the triple-quantum fil-
tered 2D correlation spectrum (Figure 6). Furthermore, the
signal pattern of the spectrum with its linking schemes AÀAÀA,
À17.7
À4.6
À9.1
À3.1
10.0
5.2
13
4
P(1), peak B
P(2), peak A
amorphous
amorphous
42.9
44.1
48
67
resonances can be revealed (see Figure 5). The two sharp
resonances, labeled A and B, are assigned to the two P sites of
the title compound, which lead to the double-quantum coher-
ences BÀB and AÀB. They can clearly be distinguished from a
broad contribution which is probably due to an amorphous
impurity. It is much broader and has its center of gravity in the
range of peak A of the crystalline title compound. The 31P peaks
were assigned according to their peak areas and according to the
different frequency of the P sites in the unit cell (Wyckoff
positions 192j (P1) and 96h (P2), Table 3). To obtain reliable
12074
dx.doi.org/10.1021/ja202159e |J. Am. Chem. Soc. 2011, 133, 12069–12078

Journal of the American Chemical Society
ARTICLE
Figure 7. 31P{1H} REDOR curve of Ba19P36O6+xN66ÀxCl8+x (x ≈
4.54) (in black). The plotted lines in green, cyan, and magenta indicate
simulated REDOR curves calculated with dipolar coupling constants
of À60, À70, and À80 Hz, respectively. The REDOR experiment proves
a small, nonstoichiometric (H atoms/unit cell <1) contamination with
hydrogen atoms, which is often observed in zeolites.38
Figure 9. Interconnected 3846812 cages in Ba19P36O6+xN66ÀxCl8+x
(x ≈ 4.54). Top: drawing with PÀN bonding and cage content (left).
Bottom: representation with only PÀP linking.
the stoichiometric occurrence of hydrogen in the structure. The
experimental REDOR curve of the deconvoluted peak B is shown
in Figure 7 together with three calculated curves for different
dipolar coupling constants of À60, À70, and À80 Hz which refer
to internuclear distances PÀH of 933, 886, and 847 pm,
respectively. There is no point in the unit cell which has such a
large distance from atom P(1). Hence, the REDOR curves prove
that small amounts of hydrogen are incorporated into the
structure; however, the observed slow dephasing effects are
indicative of substoichiometric hydrogen content (H atoms/unit
cell <1) only. Similar observations have been made previously for
many zeolites.38
Structure Description and Discussion. The crystal structure
of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) (Figure 8) consists of
Ba2+ ions, ClÀ ions, and a framework of all-side vertex-sharing
P(O/N)4 tetrahedra representing an unprecedented topology
that has been predicted by Foster and Treacy as a hypothetical
zeolite structure on the SiO2 basis.39
2+
+V[4] ÀII[2]
36
According to the IUPAC40 formula |Ba Cl8À+x|[P
O6+x
19
NÀ66IÀII[x2]]h{3[3846812]}p{3[3886]Æ100æ(8-ring)}(Fm3c), the tet-
rahedra build up 3-, 4-, and 8-rings forming 3846812 cages that
exhibit a CBU (Figure 9) that is not included in the current pool
of CBU types so far.41 The idealized solid body corresponds to
the RCSR symbol rdo-a.42 It can be derived from the Catalan
polyhedron of a rhombic dodecahedron by grinding off all
vertices. Interconnecting the novel CBUs via their 4-rings
(forming additional 3-rings) yields 8-ring channels along Æ100æ
with a free diameter of 292 pm (calculation considering r(N) =
147 pm). The coordination sequences and the vertex symbols for
the framework (RCSR symbol fuv)42 are given in the Supporting
Information (Table S4).
Figure 8. Crystal structure of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54). View
along [100] (Ba2+ black, ClÀ yellow, P(O,N)4 tetrahedra green).
AÀAÀB, AÀBÀB, and BÀBÀB corroborates the refined
structure model.
In the NPO LixH12ÀxÀy+z[P12OyN24Ày]Xz (X = Cl, Br) or the
oxonitridosodalithes M8ÀmHm[P12N18O6]Cl2 (M = Cu, Li),
hydrogen atoms are bound to the bridging nitrogen atoms forming
imide groups. The occurrence of hydrogen in Ba19P36O6+x
-
N66ÀxCl8+x (x ≈ 4.54) is difficult to prove or disprove with
diffraction techniques on an inhomogeneous mixture and be-
cause of the oxygen and nitrogen site disorder. NMR however is
rather sensitive to small amounts of hydrogen in the structure.
The 31P{1H} C-REDOR experiment can be used to investigate
The observed [Pn(N/O)n] ring sizes and their relative fre-
quency, i.e., the cycle class sequence according to Klee,43 are
listed in Table 4 (calculated with the program TOPOLAN).44 In
contrast to the sodalite framework (SOD) or the framework of
12075
dx.doi.org/10.1021/ja202159e |J. Am. Chem. Soc. 2011, 133, 12069–12078

Journal of the American Chemical Society
ARTICLE
Table 4. Cycle Class Sequence from the Framework in
Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) in Comparison with
Frameworks SOD and LTA
Pn(O,N)n-rings
3
4
5
6
7
8
9
10
Ba19P36O6+xN66ÀxCl8+x 19
4
4
0
0
8
12 31 129 388
SOD
LTA
0
0
3
4
0
0
24
69
0
0
576
360
18
56
Table 5. Selected Interatomic Distances/pm and Angles/ꢀ in
Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) (esd's in Parentheses)
Ba(1)ÀCl
311.0(12)
4 times
(partially occupied)
Ba(2)ÀCl
317.6(4), 334.2(16)
320.3(4)
3 times
8 times
8 times
7 times
8 times
Ba(3)ÀCl
Ba(1)À(O,N)
Ba(2)À(O,N)
PÀ(O,N)
308.1(5), 309.4(4)
288.6(4)À321.4(4)
159.1(6)À165.5(5)
121.5(3), 125.0(2),
125.2(3), 141.2(1)
99.0(2)À119.2(2)
PÀ(O,N)ÀP
(O,N)ÀPÀ(O,N)
12 times
Figure 10. Coordination of the Ba sites in Ba19P36O6+xN66ÀxCl8+x
(x ≈ 4.54) (Ba2+ black, ClÀ yellow (partially occupied ones are hold
transparent), (O,N) green).
LTA that comprise cubic β-(4668) and R-(4126886) cages, all ring
sizes do exist in Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54). Notably,
many 3-rings are present, although they rarely occur in zeolite
topologies (in only 15 of 194 framework types registered in the
IZA database). However, when low framework densities should
be achieved, small rings are relevant according to the theory of
Brunner and Meier.45 With a high 3-ring frequency of 19 and a
framework density of 14.87 T/1000 Å3, which is rather low
considering the typical range of zeolites (12.5À20.2 T/1000 Å3),
this theory is supported by the findings in the case of Ba19P36-
O6+xN66ÀxCl8+x (x ≈ 4.54).
saddle conformation with an intermixture of a twist chair for the
conformation in Fm3c is suggested by the values given in the
Supporting Information (section 3).
In the Rietveld refinement, constraints were included to fit
distances and angles within the tetrahedra into the usual range
observed in phosphorus nitride network structures.9,36,46 With a
certain permitted deviation the distances PÀ(O,N) vary between
159.1 and 165.5 pm and the angles (O,N)ÀPÀ(O,N) between
99.0 and 119.2ꢀ (Table 5). The angles PÀ(O,N)ÀP of the final
model range between 121.5 and 141.2ꢀ (Table 5), comparable to
The 3 4 8
8 6 12 cages in Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) are
slightly distorted from the symmetry of the regular polyhedron
(m3m). This distortion causes two different cages in the unit cell
which are crystallographically independent and thus the necessity
of the 2-fold superstructure described in Fm3c. Neglecting the
superstructure reflections and accepting just slightly higher R
values, the structure can be described in Pm3m (crystallographic
data of this basic structure are given in the Supporting Informa-
tion, Tables S1 and S2) including only one less distorted cage.
The number of independent atom sites is identical in both
models; however, P and O/N atoms occupy more special
positions in the approximate basic structure, while they occupy
more general ones in the correct space group. To quantify the
distortion of the cages in the Fm3c model and to clarify the
difference to the model in Pm3m, a conformation analysis of the
3- and 4-rings, i.e., the 6-membered P3(O,N)3 and 8-membered
P4(O,N)4 rings,46 has been performed calculating the torsion
angle sequence (TAS) as well as the displacement asymmetry
(DAP, see Supporting Information, Figure S2)47 and the pucker-
ing parameters48 with the program PARST97.49 The 6-mem-
bered rings exhibit a chair conformation in both models, which,
combined with a tiny interference of a sofa conformation, is more
pronounced in Fm3c. This can be clearly seen considering the
respective values (Supporting Information, section 3). A larger
difference is found within the 8-membered P4(O,N)4 rings.
While they are planar in Pm3m (all torsion angles are zero), a
angles in compounds like P4N4(NH)4 NH312 or SrP2N4.50 Also
3
within the corresponding hypothetical ideal SiO2 framework
generated by a distance-least-squares refinement with DLS76,51
rather small angles SiÀOÀSi down to 159.684(2)ꢀ are existent.
This might be the reason why this framework has not been
realized in an oxidic system as of yet.
We assume a statistic O,N-distribution in the P/O/N network
of Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54). As there are no predes-
tined positions for O or N and such statistic O,N-distributions
were already observed in other P/O/N TX2 networks,9,6,7,52 this
assumption is reasonable. Possible O,N ordering in Ba19P36-
O6+xN66ÀxCl8+x (x ≈ 4.54) could only be analyzed by neutron
diffraction. From the refined sum formula Ba19P36O10.54
-
N61.46Cl12.54 that is confirmed by semiquantitative EDX analysis
within the accuracy of the method (calcd: Ba:Cl = 1.5, Ba:P = 0.5,
P:(O,N) = 0.5, P:Cl = 2.9, N:O = 5.8; exptl: Ba:Cl = 1.3, Ba:P =
0.6, P:(O,N) = 0.4, P:Cl = 2.4, N:O = 4.4), on average 56% PN4
and 44% PON3 tetrahedra are present.
The framework in Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) pro-
vides a “free” volume of 4484.3 Å3 (23.2% of the cell volume).
Within this volume, which comprises the 8-ring channels and the
interior of the cages, the extra-framework ions Ba2+ and ClÀ are
located. One Ba2+ site lies directly in the center of the cages
12076
dx.doi.org/10.1021/ja202159e |J. Am. Chem. Soc. 2011, 133, 12069–12078

Journal of the American Chemical Society
ARTICLE
coordinated by a cube of eight ClÀ ions at a distance of 320.3 pm
(Figure 10, Ba(3)), corresponding to the content of one unit cell
of the CsCl structure type. Around the BaCl86À units, 12 more
Ba2+ ions are located within each cage. These ions are coordi-
nated 10-fold by seven (O,N) atoms (288.6À321.4 pm) of the
cage and three ClÀ ions (317.6 and 334.2 pm; Figure 10, Ba(2)).
The Ba2+ ions along the 8-ring channels are arranged in pairs with
an intrapair distance of 566.5 pm and ClÀ ions in between
(Figure 10, Ba(1)). The four ClÀ positions are staggered from
pair to pair (cf. Figure 8) and have, due to their partial occupancy
(about 40%), with 311.0 pm the shortest BaÀCl distance in
Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54). Furthermore, the Ba2+ ions
involved in the pairs which have an interpair distance of 1342.7
pm are surrounded each by eight (O,N) atoms with distances
between 308.1 and 309.4 pm. All BaÀCl and BaÀ(O,N)
contacts, summarized in Table 5, range in the sum of the
respective ionic radii.53,54 Although the cages and channels are
quite crowded with Ba2+ and ClÀ, there is still free space (about
38 Å3) between the Ba2+ ions that are situated in the periphery of
the cage interior in a distance of around 200 pm to the bridging
atoms O,N. Replacing ClÀ and BrÀ ions can also be partially
introduced by adding NH4Br to the starting material mixture. A
replacement of ClÀ by BrÀ could be proved obtaining cubic
microcrystals (Figure S3, Supporting Information) that contain
the elements Ba, P, O, N, Cl, and Br (EDX analysis) and powder
diffraction data where all reflections move to lower 2θ-values
compared to the pure ClÀ containing compound (a =
2691.06(3) pm). No evidence was observed for incorporation
of IÀ or cations other than Ba2+.
the multicomponent reactant system with phosphorus triamides,
metal halides, and ammonium halides. A comprehensive inves-
tigation of the role of the NH3 partial pressure as well as the
templating effect of Ba2+ and ClÀ would allow one to purpose-
fully optimize the synthesis. Other soft and large ions like Pb2+ or
Sn2+ might as well be useful to establish a systematic route for the
synthesis of nitridic zeolites and permit novel framework types
with low framework densities and thus exploit the full potential of
the nitride approach within zeolite chemistry.
’ ASSOCIATED CONTENT
S
Supporting Information. Simulated electron diffraction
b
pattern compared with the experimental pattern of direction
[100], crystallographic data in space group Pm3m, symmetry of
6-membered P3(O,N)3 and 8-membered P4(O,N)4 rings, topol-
ogy analysis (coordination sequences and vertex symbols), and an
SEMimageoftheisostructural phasecontainingBrÀ. This material
is available free of charge via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
wolfgang.schnick@uni-muenchen.de
’ ACKNOWLEDGMENT
We gratefully acknowledge financial support that was granted
by the Deutsche Forschungsgemeinschaft (DFG) and Fonds der
Chemischen Industrie (FCI) (scholarship for S. J. Sedlmaier,
Emmy Noether funding for J. Schmedt auf der G€unne and J.
Weber). Additionally we thank ESRF, Grenoble, France for
supplying beamtime and Dr. A. Fitch for his assistance at ID31.
Our special thanks go to Dr. C. B€arlocher for the DLS76
calculations and valuable discussions. We also thank C. Minke
for numerous SEM images and EDX measurements.
’ CONCLUSION
With the novel oxonitridophosphate Ba19P36O6+xN66ÀxCl8+x
(x ≈ 4.54), an unprecedented zeolite-like framework including a
new CBU unit has been elucidated. Its discovery and its structure
elucidation were highly challenging, as from a complex synthesis
system only microcrystalline powder samples with side phases
were available. However, applying a combination of high-resolu-
tion synchrotron powder diffraction and electron microscopic
methods (SAED and STEM), ambiguous data could be clarified,
and cubic (Fm3c (no. 226), a = 2685.41(3) pm) Ba19P36-
O6+xN66ÀxCl8+x (x ≈ 4.54) could be established as a compound
exhibiting a novel all-side vertex-sharing P(O,N)4 tetrahedra
topology that has been predicted but not observed as of yet.
Confirmed and complemented with solid-state NMR spectros-
copy, the structure model comprises Ba2+ and ClÀ ions which are
’ REFERENCES
(1) Cheetham, A. K.; Fꢀerey, G.; Loiseau, T. Angew. Chem. 1999,
111, 3466–3492. Angew. Chem., Int. Ed. 1999, 38, 3268–3292.
(2) Natarajan, S.; Mandal, S. Angew. Chem. 2008, 120, 4876–4907.
Angew. Chem., Int. Ed. 2008, 47, 4798–4828.
(3) Wilson, S. T.; Lok, B. M.; Messina, C. A.; Cannan, T. R.;
Flanigen, E. M. J. Am. Chem. Soc. 1982, 104, 1146–1147.
(4) (a) Guth, J. L.; Kessler, H.; Wey, R. Stud. Surf. Sci. Catal. 1986,
28, 121–128. Estermann, M.; McCusker, L. B.; Baerlocher, C.;
Merrouche, M.; Kessler, H. Nature 1991, 352, 320–323.
(5) Bowes, C. L.; Ozin, G. A. Adv. Mater. 1996, 8, 13–28.
(6) (a) Schnick, W.; L€ucke, J. Angew. Chem. 1992, 104, 208–209.
Angew. Chem., Int. Ed. Engl. 1992, 31, 213–215. (b) Schnick, W.; L€ucke, J.
Z. Anorg. Allg. Chem. 1994, 620, 2014–2019. (c) Schnick, W.; Stock, N.;
L€ucke, J.; Volkmann, M.; Jansen, M. Z. Anorg. Allg. Chem. 1995,
621, 987–992. (d) Wester, F.; Schnick, W. Z. Anorg. Allg. Chem. 1996,
622, 1281–1286.
incorporated in interlinked 3 4 8
8 6 12 cages and in 8-ring channels
with a free diameter of 292 pm. If this extra-framework material
could be (partially) removed or exchanged by smaller ions, which
will be the subject of future investigations, the title compound
may be a representative for a small-pore zeolite that can usually
be used in pressure swing adsorption (PSA) separation
processes.55À57 This or other applications seem to be desirable
as Ba19P36O6+xN66ÀxCl8+x (x ≈ 4.54) has a surprising thermal
stability up to at least 1100 ꢀC, a temperature where many
common zeolites are already decomposed. With this stability,
nitride materials would be interesting for catalytic processes that
act at high temperatures.
Furtherimprovement anddevelopmentof thecomplexsynthe-
sis system must be another future goal. After the (oxo-)nitrido-
sodalites and the NPO type materials, Ba19P36O6+xN66ÀxCl8+x
(x ≈ 4.54) is another important example for the use of
(7) Stock, N.; Irran, E.; Schnick, W. Chem.—Eur. J. 1998,
4, 1822–1828.
(8) Huppertz, H.; Schnick, W. Angew. Chem. 1997, 109, 2765–2767.
Huppertz, H.; Schnick, W. Angew. Chem., Int. Ed. Engl. 1997,
36, 2651–2652.
(9) (a) Correll, S.; Oeckler, O.; Stock, N.; Schnick, W. Angew. Chem.
2003, 115, 3674–3677. Angew. Chem., Int. Ed. 2003, 42, 3549–3552.
(b) Correll, S.; Stock, N.; Oeckler, O.; Senker, J.; Nilges, T.; Schnick, W.
Z. Anorg. Allg. Chem. 2004, 630, 2205–2217.
12077
dx.doi.org/10.1021/ja202159e |J. Am. Chem. Soc. 2011, 133, 12069–12078

Journal of the American Chemical Society
ARTICLE
(10) Barnes, A. J. D.; Prior, T. J.; Francesconi, M. G. Chem. Commun.
2007, 4638–4640.
(11) Pagano, S.; Oeckler, O.; Schr€oder, T.; Schnick, W. Eur. J. Inorg.
Chem. 2009, 2678–2683.
(12) Karau, F.; Schnick, W. Angew. Chem. 2006, 118, 4617–4620.
Angew. Chem., Int. Ed. 2006, 45, 4505–4508.
(48) Cremer, D.; Pople, J. A. J. Am. Chem. Soc. 1975, 97, 1354–1358.
(49) Nardelli, M. PARST97, Univ. Parma: Italy, 1997.
(50) Karau, F. W.; Seyfarth, L.; Oeckler, O.; Senker, J.; Landskron,
K.; Schnick, W. Chem.—Eur. J. 2007, 13, 6841–6852.
(51) Baerlocher, Ch.; Hepp, A.; Meier, W. M. Distance Least Squares
Refinement Program, DLS-76; ETH: Z€urich, 1977.
(13) Pouchard, M. Nature 2006, 442, 878–879.
(52) Lꢀeger, J. M.; Haines, J.; Chateau, C.; Bocquillon, G.; Schmidt,
M. W.; Hull, S.; Gorelli, F.; Lesauze, A.; Marchand, R. Phys. Chem. Miner.
2001, 28, 388–398.
(53) Shannon, R. D. Acta Crystallogr., Sect. A: Found. Crystallogr.
1976, 32, 751–767.
(54) Baur, W. H. Crystallogr. Rev. 1987, 1, 59–83.
(55) Zhu, W.; Kapteijn, F.; Moulijn, J. A.; den Exter, M. C.; Jansen,
J. C. Langmuir 2000, 16, 3322–3329.
(14) Klement, R.; Koch, O. Chem. Ber. 1954, 87, 333–340.
(15) Schnick, W. Z. Naturforsch., B: J. Chem. Sci. 1989, 44, 942–945.
(16) Correll, S. Ph.D. Thesis, Univ. Munich (LMU) 2006.
(17) Horstmann, S.; Schnick, W. Z. Naturforsch., B: J. Chem. Sci.
1996, 51, 127–132.
(18) (a) Sowerby, D. B.; Audrieth, L. F. Chem. Ber. 1961,
94, 2670–2675. (b) Golinski, F.; Jacobs, H. Z. Anorg. Allg. Chem.
1994, 620, 965–968.
(56) Olson, D. H. U.S. Patent 6 488 741, 2002.
(19) Stock, N.; Schnick, W. Z. Naturforsch., B: J. Chem. Sci. 1996,
51, 1079–1083.
(57) Olson, D. H.; Camblor, M. A.; Villaescusa, L. A.; Kuehl, G. H.
Microporous Mesoporous Mater. 2004, 67, 27–33.
(20) Hodeau, J.-L.; Bordet, P.; Anne, M.; Prat, A.; Fitch, A. N.;
Dooryhee, E.; Vaughan, G.; Freund, A. K. Proc. SPIE 1998,
3448, 353–361.
(21) Coelho, A. A. TOPAS-Academic, Version 4.1; Coelho Software:
Brisbane, 2007.
(22) Coelho, A. A. J. Appl. Crystallogr. 2003, 36, 86–95.
(23) Oszlꢀanyi, G.; S€uto, A. Acta Crystallogr., Sect. A: Found. Crystal-
logr. 2004, 60, 134–141.
(24) Bergmann, J.; Kleeberg, R.; Haase, A.; Breidenstein, B. Mater.
Sci. Forum 2000, 347À349, 303–308.
(25) Sabine, T. M.; Hunter, B. A.; Sabine, W. R.; Ball, C. J. J. Appl.
Crystallogr. 1998, 31, 47–51.
(26) Hill, R. J.; Madsen, I. C. J. Appl. Crystallogr. 1984, 17, 297–306.
(27) Hill, R. J.; Flack, H. D. J. Appl. Crystallogr. 1987, 20, 356–361.
(28) Jouneau, P.-H.; Stadelmann, P. Electron Microscopy Image
Simulation (EMS On Line); Centre Interdꢀepartemental de Microscopie
Electronique, EPFL: Lausanne, 1998: http://cecm.insa-lyon.fr/CIOLS/
crystal1.pl.
(29) Harris, R. K.; Becker, E. D.; Cabral de Menezes, S. M.;
Goodfellow, R.; Granger, P. Solid State NMR 2002, 22, 458–483.
(30) Hohwy, M.; Jakobsen, H. J.; Edꢀen, M.; Levitt, M. H.; Nielsen,
N. C. J. Chem. Phys. 1998, 108, 2686–2694.
(31) Shantz, D. F.; Schmedt auf der G€unne, J.; Koller, H.; Lobo, R. F.
J. Am. Chem. Soc. 2000, 122, 6659–6663.
(32) Chan, J. C. C.; Eckert, H. J. Chem. Phys. 2001, 115, 6095–6105.
(33) Roming, M.; Feldmann, C.; Avadhut, Y. S.; Schmedt auf der
G€unne, J. Chem. Mater. 2008, 20, 5787–5795.
(34) Tessier, F.; Navrotsky, A.; Le Sauze, A.; Marchand, R. Chem.
Mater. 2000, 12, 148–154.
(35) Sedlmaier, S. J.; Ebersp€acher, M.; Schnick, W. Z. Anorg. Allg.
Chem. 2011, 637, 362–367.
(36) (a) Landskron, K. Ph.D. Thesis, Univ. Munich (LMU), 2001.
(b) Karau, F. Ph.D. Thesis, Univ. Munich (LMU), 2007.
(37) Sedlmaier, S. J.; Schmedt auf der G€unne, J.; Schnick, W. Dalton
Trans. 2009, 4081–4084.
(38) Brunner, E. J. Mol. Struct. 1995, 355, 61–85.
(39) Foster, M. D.; Treacy, M. M. J. A Database of Hypothetical Zeolite
Structures: http://www.hypotheticalzeolites.net, accessed June 2010.
(40) McCusker, L. B.; Liebau, F.; Engelhardt, G. Pure Appl. Chem.
2001, 73, 381–394.
(41) Baerlocher, Ch.; McCusker, L. B. Database of Zeolites Structures:
http://www.iza-structure.org/databases/, accessed November 2010.
(42) O’Keeffe, M.; Peskov, M. A.; Ramsden, S. J.; Yaghi, O. M. Acc.
Chem. Res. 2008, 41, 1782–1789.
(43) Klee, W. E. Z. Kristallogr. 1987, 179, 67–76.
(44) Thimm, G.; Schumacher, S.; Uhr, W.; Klee, W. E. TOPOLAN,
Topological Analysis of Crystal Structures; Univ. Karlsruhe: Germany, 1993.
(45) Brunner, G. O.; Meier, W. M. Nature 1989, 337, 146–147.
(46) Stock, N. Ph.D. Thesis, Univ. Bayreuth, 1998.
(47) Nardelli, M. Acta Crystallogr., Sect. C: Cryst. Struct. Commun.
1983, 39, 1141–1142.
12078
dx.doi.org/10.1021/ja202159e |J. Am. Chem. Soc. 2011, 133, 12069–12078