X-ray diffraction, 133Cs and 1H NMR and thermal studies of CdZrCs1.5(H3O)0.5(C 2O4)4·xH2O displaying Cs and water dynamic behavior

Article abstract of DOI:10.1016/j.jpcs.2004.01.008

A novel mixed cadmium zirconium cesium oxalate with an open architecture has been synthesized from precipitation methods at room pressure. It crystallizes with an hexagonal symmetry, space group P3₁12 (no. 151), a=9.105(5)?, c=23.656(5)?, V=1698(1)?³ and Z=3. The structure displays a [CdZr(C₂O₄)₄] ^2- helicoidal framework built from CdO₈ and ZrO ₈ square-based antiprisms connected through bichelating oxalates, which generates channels along different directions. Cesium cations, hydronium ions and water molecules are located inside the voids of the anionic framework. They exhibit a dynamic disorder which has been further investigated by ¹H and ¹³³Cs solid-state NMR. Moreover a phase transition depending both upon ambient temperature and water vapor pressure was evidenced for the title compound. The thermal decomposition has been studied in situ by temperature-dependent X-ray diffraction and thermogravimetry. The final product is a mixture of cadmium oxide, zirconium oxide and cesium carbonate.

Full text of DOI:10.1016/j.jpcs.2004.01.008

Journal of Physics and Chemistry of Solids 65 (2004) 1213–1221
www.elsevier.com/locate/jpcs
133
1
X-ray diffraction, Cs and H NMR and thermal studies
of CdZrCs1.5(H3O)0.5(C2O4)4·xH2O displaying Cs and water
dynamic behavior
a
E. Jeanneau , M. Le Floch , B. Bureau , N. Audebrand , D. Louer
b
b
a
a,
*
¨
a
´ ´
Laboratoire de Chimie du Solide et Inorganique Moleculaire (UMR 6511 CNRS), Institut de Chimie, Universite de Rennes,
´ ´
Avenue du General Leclerc, 35042 Rennes cedex, France
b
´ ´ ´ ´
Laboratoire Verres et Ceramiques (UMR 6512 CNRS), Institut de Chimie, Universite de Rennes, Avenue du General Leclerc,
35042 Rennes cedex, France
Received 18 September 2003; revised 7 January 2004; accepted 21 January 2004;
Available online 4 March 2004
Abstract
A novel mixed cadmium zirconium cesium oxalate with an open architecture has been synthesized from precipitation methods at room
3
˚
˚
˚
pressure. It crystallizes with an hexagonal symmetry, space group P3112 (no. 151), a ¼ 9:105ð5Þ A, c ¼ 23:656ð5Þ A, V ¼ 1698ð1Þ A and
Z ¼ 3: The structure displays a [CdZr(C2O4)4]22 helicoidal framework built from CdO8 and ZrO8 square-based antiprisms connected
through bichelating oxalates, which generates channels along different directions. Cesium cations, hydronium ions and water molecules are
located inside the voids of the anionic framework. They exhibit a dynamic disorder which has been further investigated by H and 133Cs
1
solid-state NMR. Moreover a phase transition depending both upon ambient temperature and water vapor pressure was evidenced for the title
compound. The thermal decomposition has been studied in situ by temperature-dependent X-ray diffraction and thermogravimetry. The final
product is a mixture of cadmium oxide, zirconium oxide and cesium carbonate.
q 2004 Elsevier Ltd. All rights reserved.
Keywords: C. X-ray diffraction; D. Crystal structure; D. Nuclear magnetic resonance; D. Phase transitions
1. Introduction
examples with counter-cations adopting also an eight-fold
coordinationareCd2Zr(C2O4)4·6H2O[1]andCdZrK2(C2O4)4·
8H2O [2], in which the anionic framework is counter-
balanced by one Cd2þ or two Kþ cations, respectively,
located in the voids of the structures. The topology of the
anionic framework based on square antiprism polyhedra
presents an hexagonal symmetry and is built from
interconnected helical chains exhibiting a sequence similar
to that found in the zigzag chains. The cations located inside
the channels parallel to the c-axis counterbalance the
anionic framework. This structure has been reported for
three compounds with inserted (metallic and organic)
cations of different charges, i.e. CdZr(NH4)2(C2O4)4·3.9-
H2O [3], CdZr(C2N2H10)(C2O4)4·4.4H2O [3] and
CdZrNa2(C2O4)4·8.5H2O [4]. For the last material, a
dynamic disorder of the cations and water molecules
located inside the channels was pointed out from single-
crystal X-ray diffraction and 1H and 23Na solid-state NMR.
These properties have motivated the investigation of new
Recent studies on mixed cadmium–zirconium oxalates
have revealed the existence of a family of phases with
structural arrangements based on MO8 (M ¼ Cd, Zr)
polyhedra connected by bichelating oxalate groups [1–4].
Two open structure frameworks have been pointed out in the
family [2,3]. Both of them are built from [CdZr(C2O4)4]22
anionic groups forming two different frameworks with a
topology originating from the two frequent configurations in
the stereochemistry of eight-fold coordination, i.e. the
dodecahedron and the square antiprism [5,6]. With dodeca-
hedra, the structure is built from a net of perpendicular zigzag
chains with a (–Cd-oxalate–Zr-oxalate–)1 sequence
and exhibits tunnels with square cross-sections. Representative
*
Corresponding author. Tel.: þ33-223-23-62-48; fax: þ33-2-99-38-
34-87.
¨
E-mail address: daniel.louer@univ-rennes1.fr (D. Louer).
0022-3697/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jpcs.2004.01.008

1214
E. Jeanneau et al. / Journal of Physics and Chemistry of Solids 65 (2004) 1213–1221
phases involving an alkaline element. The present study
deals with the new cesium-based compound, its crystal
structure determination from X-ray diffraction, its thermal
during the thermogravimetric analysis presented in this
study.
1
behavior and the use of 133Cs and H solid-state NMR to
2.2. Single-crystal data collection
investigate the dynamic disorder inside the channels of the
structure framework.
A suitable single crystal was carefully selected. Intensity
data were collected on a four-circle Nonius KappaCCD
diffractometer equipped with a CCD area detector through
the program COLLECT [7]. Lorentz-polarisation factor
correction, peak integration and background determination
were carried out with the program DENZO [8]. Frame
scaling and unit-cell parameters refinement were made with
the program SCALEPACK [8]. Numerical absorption
correction was performed by modeling the crystal faces
using NUMABS [9], included in the WinGX program
platform [10]. The resulting set of hkl reflections was used
for structure refinement. Crystallographic data and details
on data collection and refinement are listed in Table 1.
Structure drawings were carried out with Diamond 2.1e,
supplied by Crystal Impact [11].
2. Experimental section
2.1. Synthesis and primary characterization
Crystals of the ternary cadmium zirconium cesium
oxalate were obtained through a mild chemistry route. The
title compound was synthesized from a mixture of 0.5 g of
cadmium nitrate Cd(NO3)2·4H2O (Merck), 0.37 g of
zirconium hydroxynitrate (Alpha) and 1.12 g cesium nitrate
CsNO3 (Aldrich), dissolved in 50 ml of deionized water. A
white precipitate was formed from the titration of this
solution by an aqueous solution of oxalic acid (0.1 mol l21
)
in excess. The solution was then heated to ,333 K and
,3 ml of concentrated nitric acid were slowly poured in,
under constant stirring, until complete dissolution of the
previous precipitate. The evaporation of the solution at
room temperature led to the formation of small transparent
bipyramidal crystals. These crystals were subsequently
washed with water and ethanol, and dried under air.
2.3. Collection of X-ray powder diffraction data
X-ray powder diffraction data were obtained with a
Siemens D500 diffractometer with the parafocusing Bragg–
Brentano geometry, using monochromatic Cu Ka1
˚
radiation (l ¼ 1:5406 A) selected with an incident beam
Even though the structure framework topology of this
new cesium phase is analogous to that of previously
reported compounds CdZrA(C2O4)4·x H2O [A ¼ Na2,
(NH4)2, (C2N2H10)] [3,4], the title compound possesses
interesting features. One of these concerns its non-
stoichiometry in cesium which is to be associated with
cationic exchange properties. The proposed chemical
formula of the compound studied here is CdZrCs1.5
(H3O)0.5(C2O4)4·x H2O. This formulation was derived
from energy dispersive spectroscopy performed using a
JSM 6400 spectrometer equipped with an Oxford Link Isis
analyzer and atomic absorption spectroscopy analysis with a
VARIAN SpetrAA 10 plus. The mean value, found from
nine energy dispersive spectroscopy analyses on Cd, Zr and
Cs, is 1.5(3) cesium atom per formula unit. This value was
confirmed by eight Cd and Cs atomic absorption analyses,
which gave a number of 1.4(1) cesium atom per formula
unit. The structure solution from single-crystal X-ray
diffraction data did not allow to confirm the amount of
cesium atoms because of a disorder phenomenon, which is
discussed in this study. The number of cesium atoms
determined does not counterbalance the anionic charge of
the oxalate anions. It is thus necessary to consider that other
positive charges are present. As expected from the
synthesis, no other heavy atom ðZ . 8Þ was detected by
EDS. Consequently, according to the acidic solution used in
the preparation, it is reasonable to suggest the presence of
H3Oþ ions. Moreover, the content in hydronium ions,
together with Csþ, is consistent with weight losses observed
curved-crystal germanium monochromator [12]. For pattern
Table 1
Crystallographic data and structure refinement parameters for the anionic
framework [CdZr(C2O4)4]22 in the compound CdZrCs1.5(H3O)0.5(C2O4)4·
x H2O
Empirical formula
Crystal system
Space group
[CdZr(C2O4)4]22
Hexagonal
P3112 (No. 151)
0.05 £ 0.05 £ 0.05
9.105(5)
Crystal size (mm)
˚
a (A)
˚
c (A)
23.656(5)
1698(1)
3
˚
V (A )
Z
3
˚
l (Mo Ka), A
0.71073
u range (8)
Index ranges
2.58–36.72
215 # h # 15;
212 # k # 12;
239 # l # 39
5451
Unique data
Observed data ðI . 2sðIÞÞ
R1 ðI . 2sðIÞÞ
R1 (All)
2864
0.048
0.091
vR2 ðI . 2sðIÞÞ
vR2 (All)
Refinement method
0.141
0.157
Full-matrix
least-squares on lF2l
0.973
Goodness of fit
No. of variables
121
Largest difference map peak
23
0.709 and 21.648
˚
and hole (e A
)

E. Jeanneau et al. / Journal of Physics and Chemistry of Solids 65 (2004) 1213–1221
1215
indexing, the extraction of the peak positions was carried
2.5. Thermal analysis
out with the Socabim program PROFILE, which is a part of
the DIFFRACT-AT software package supplied by Bruker
AXS. Pattern indexing was performed with the program
DICVOL91 [13] and full-pattern matching with FULL-
PROF[14], availableinthesoftwarepackageWinPLOTR [15].
Temperature-dependent X-ray diffraction (TDXD) was
performed under dynamic air with a powder diffractometer
combining the curved-position-sensitive detector from
INEL (CPS 120) and a high-temperature attachment from
Rigaku. The detector was used in a semi-focusing arrange-
ment by reflection (Cu Ka1 radiation) as described else-
where [16].
Thermogravimetric analyses (TG) were carried out with
a Rigaku thermoflex instrument for runs under airflow. The
powdered samples were spread evenly in large platinum
crucibles to avoid mass effects.
3. Results and discussion
3.1. Structure solution
The structure of CdZrCs1.5(H3O)0.5(C2O4)4·x H2O was
solved in the hexagonal symmetry, space group P3.
The cadmium and zirconium atoms were localized using
the direct methods through the program SIR97 [19]. The
remaining atoms, constitutive of the anionic framework
[CdZr(C2O4)4]22, were localized by successive Fourier
difference map analysis with the help of SHELXL97
program [20]. Once all the atoms from the anionic
framework were localized, the remaining maximum
2.4. NMR measurements
The spectra were recorded at room temperature on a ASX
300 Bruker spectrometer operating at 300 MHz for 1H ðI ¼
1=2Þ and 39.35 MHz for 133Cs ðI ¼ 7=2Þ with a 4 mm MAS
probe spinning up to 15 kHz. A single pulse sequence was
used to acquire the proton signal. The pulse duration was
3 ms and the recycle delay was set to 1 s after having
checked that no contribution was favored due to T1 effect.
For 133Cs, the Hahn spin echo sequence was applied for
refocusing magnetization lost in the dead time. The first
pulse length was chosen equal to 1 ms (2 ms for the second
pulse) in order to fulfil the condition nrf , nQ that avoids
any line distortion due to quadrupolar effects [17]. CsCl
dissolved in water was used as reference (0 ppm). The
simulations of the experimental spectra were performed
using the Dm2000nt version of the Winfit software [18].
The definitions of the measured quadrupolar parameters are
the following. VXX; VYY and VZZ are the principal
components of the electric field gradient tensor, with the
condition lVZZl $ lVYY l $ lVXXl and VZZ ¼ eq: The quad-
rupolar coupling constant is written
23
˚
residual density peaks were weaker than 6.57 e A and
could thus not be attributed to a cesium atom. Fourier
difference maps, made by including the atoms of the
framework, revealed a completely delocalized electronic
density inside the channels (Fig. 1a and b). A low
temperature X-ray diffraction study was attempted but
failed due to the transformation of the single crystal into a
polycrystalline pseudomorph under the nitrogen flux.
The refinement of the anionic framework was then
performed by ignoring the contribution of the disordered
part of the structure. For this, the region containing the
disordered electronic density was identified by considering
the Van der Waals radii of the atoms constitutive of the
ordered part. The contribution of this region to the total
structure factor was calculated via a discrete Fourier
transformation, and incorporated in a further least-squares
refinement. These two procedures were carried out with
the program SQUEEZE [21] included in the WinGX
platform [10]. It resulted in a significant decrease of the R
values and a convergence of the refinement. Moreover, this
procedure gives an electron count over the disordered region
of 511, which is in agreement with the presence of 1.5
cesium atoms, 0.5 H3Oþ and about eight water molecules
per formula unit (total of 504 electron per unit cell). This
hydration content is consistent with that determined for the
analogous phase Cd/Zr/Na [4]. The analysis of the
symmetry elements of the ordered part of the structure
done with ADDSYM [22,23] showed that the structure
framework was in agreement with space group P3112. The
atomic coordinates of the framework, together with the
equivalent isotropic displacement parameters, are given in
Table 2 according to the space group P3112. The mean
distances are reported in Table 3 according to Hoard and
Silverton’s formalism [5].
e2qQ
h
eQVZZ
¼
h
CQ ¼
where Q is the quadrupolar moment (23 £ 10231 m2 for
133Cs) and the asymmetry parameter is given by hQ ¼
ðVXX 2 VYY Þ=VZZ: The chemical shift is defined as dii ¼
sref 2 sii; where sii ði ¼ x; y; zÞ are the principal com-
ponents of the shielding tensor. The isotropic chemical shift
diso; the anisotropy dc:s: and asymmetry hc:s: parameters used
for the reconstruction are defined as follows:
1
diso
¼
¼
ðdXX þ dYY þ dZZÞ; dc:s: ¼ dZZ 2 diso and hc:s:
3
dYY 2 dXX
:
diso

1216
E. Jeanneau et al. / Journal of Physics and Chemistry of Solids 65 (2004) 1213–1221
Fig. 1. Views of the [CdZr(C2O4)4]22 anionic framework, together with the electronic density contained inside the channels (isosurface at 1.5 e A23), along (a)
the c-axis and (b) the a-axis.
˚
3.2. Structure description
distance (2.21(3) A) is close to the theoretical distance
calculated by the bond valence method [25] for an eight-fold
˚
coordinated zirconium atom (2.18 A). The mean Cd–O
˚
distance (2.5(1) A] is also close to the calculated value for an
˚
eight-fold coordinated cadmium atom (2.41 A). It can be
The compound CdZrCs1.5(H3O)0.5(C2O4)4·x H2O can be
considered as made of a three-dimensional [CdZr(C2O4)4]22
anionic framework with inserted disordered counter-cations
(Csþ and H3Oþ). The anionic framework is built up from
cadmium and zirconium cations linked through oxalate
groups and is isotypic with those already thoroughly
described for the compounds CdZr(NH4)2(C2O4)4·3.9H2O,
CdZr(C2N2H10)(C2O4)4·4.4H2O [3] and CdZrNa2(C2O4)4·
8.5H2O [4]. Indeed the eight-fold metal atoms arrange
themselves in interconnected helical wires which displays
noted that half the Cd–O distances are longer than expected,
whereas the remaining ones are shorter.
Two oxalate groups, linking the cadmium and zirconium
atoms in a bidentate way, are present in the asymmetric unit.
The mean distances and angles within these oxalate groups
are reported in Table 3. They are close to the values reported
by Hahn [26] for oxalate-based compounds, i.e. 1.24 and
˚
tunnels with a circular section along c (Ø , 5 A) (Fig. 1a)
˚
1.55 A, 117 and 1258 for the C–O and C–C bonds, and
˚
˚
and with elliptic (,5 £ 10 A) and square (,5 A) sections
O–C–C and O–C–O angles, respectively. The two oxalate
groups are nearly planar with a maximum atomic deviation
˚
from mean plane of 0.04 A.
along a; b and [110] (Fig. 1b). The parameters of an helix,
˚
as defined in Ref. [3], are its period P ¼ c ¼ 23; 656 A and
˚
its radius r ¼ a=2 ¼ 4:553 A. The total volume available
3
for the disordered cesium and hydronium cations is 802 A ,
˚
Table 2
2
Atomic coordinates and atomic displacement parameters (A ) for the atoms
which corresponds to approximately 47% of the unit-cell
volume. By analogy with the related compounds cited
above, the counter-cations can be supposed to be inside the
circular channels along the 31 axis. Water molecules are
also present in the different types of channels. The
structure study clearly reveals a disorder problem inside
the tunnels. To further elucidate this feature, a local probe
technique must be used. Solid-state NMR spectroscopy has
been employed here to understand this phenomenon. The
results are reported below.
˚
involved in the framework of CdZrCs1.5(H3O)0.5(C2O4)4·xH2O, Ueq:
j Uijapi apj ai·aj
¼
P P
i
ð1=3Þ
2
Ueq: (A )
˚
Atom
x=a
y=b
z=c
Cd
Zr
5/6
1/6
1/3
0.0266(1)
0.0198(1)
0.0359(7)
0.049(1)
1/3
1/6
1/2
O1
O2
O5
C1
C2
O4
O3
O8
O7
O6
C3
C4
0.5574(4)
0.7585(5)
1.0726(4)
0.6334(6)
0.5400(5)
0.4184(4)
0.5954(4)
0.1110(5)
0.9067(4)
0.2510(4)
1.1244(5)
0.0304(6)
0.3431(4)
0.3652(5)
0.3250(4)
0.2854(8)
0.0853(7)
0.0178(4)
0.0091(5)
20.0103(5)
20.0344(5)
0.3184(4)
0.2475(7)
0.0478(7)
0.4475(2)
0.3906(2)
0.3896(1)
0.4177(2)
0.4193(1)
0.4524(1)
0.3891(1)
0.4480(2)
0.3901(2)
0.4527(1)
0.4188(1)
0.4184(1)
0.0333(7)
0.028(1)
The geometry of the ZrO8 and CdO8 polyhedra was
determined using Haigh’s criterion [24], based on the
analysis of O–M–O angles. It shows that both polyhedra
are square-based antiprisms. The mean planes of the square
bases have 0.01 and 0.068 maximum atomic deviations from
planarity for CdO8 and ZrO8, respectively. Moreover, these
bases are nearly parallel with a 0.22 and 0.368 angle between
the two mean planes for CdO8 and ZrO8. The mean Zr–O
0.0270(9)
0.0309(7)
0.0386(8)
0.0402(8)
0.0432(9)
0.0320(7)
0.0267(9)
0.0277(9)

E. Jeanneau et al. / Journal of Physics and Chemistry of Solids 65 (2004) 1213–1221
1217
Table 3
intensities ratio of the two phases. This feature ruled out
the presence of an impurity and suggested the existence of a
subtle phase transformation around ambient atmospheric
conditions. Taking into account the sensitivity of some other
oxalate-based compounds such as YK(C2O4)2·4H2O [27] or
Cd2Zr(C2O4)2·(4 þ x)H2O [1] towards water vapor pressure
and temperature, it can be thought that the phase transition
in the cesium compound could be due to a variation of the
water content in the channels. The two phases were
successfully isolated and characterized by X-ray powder
diffraction at room temperature (295 K) with a humidity of
37% for the low hydrated phase (LH) (Fig. 2a) and 65% for
the high hydrated phase (HH) (Fig. 2b). For each of these
two phases, the unit-cell parameters were refined using the
full-pattern matching option included in the FULLPROF
program suite. Observed line profiles were represented by a
pseudo-Voigt function. The refinement also included the
zero shift, the three FWHM parameters defining the Caglioti
function, two variables describing the angular dependence
of the profile shape factor and two parameters for peak
asymmetry correction in the 10–408 ð2uÞ angular range. The
˚
Selected interatomic distances (A) and bond angles (8) in the structure
refinement of CdZrCs1.5(H3O)0.5(C2O4)4·x H2O
ZrO8 antiprism
M–O
Zr–O1, O1(ii)
Zr–O6, O6(ii)
2.236(4)
2.182(3)
Zr–O4, O4(ii)
Zr–O8, O8(ii)
2.180(3)
2.226(4)
l
s
O1–O1(ii)
O1(ii)–O4
O4–O6(ii)
O6(ii)–O8
O8–O8(ii)
O8(ii)–O6
O6–O4(ii)
O4(ii)–O1
2.745(7)
2.853(4)
2.604(4)
2.843(4)
2.737(7)
2.843(4)
2.604(4)
2.853(4)
O1–O4
2.577(5)
2.682(5)
2.604(5)
2.687(5)
2.687(5)
2.604(5)
2.682(5)
2.577(5)
O4–O8
O8–O6
O6–O1
O1(ii)–O6(ii)
O6(ii)–O8(ii)
O8(ii)–O4(ii)
O4(ii)–O1(ii)
CdO8 antiprism
M–O
Cd–O2, O2(i)
Cd–O5, O5(i)
2.607(4)
2.336(3)
Cd–O3, O3(i)
Cd–O7, O7(i)
2.320(3)
2.613(4)
l
s
O2–O3(i)
O3(i)–O5
O5–O7(i)
O7(i)–O7
O7–O5(i)
O5(i)–O3
O3–O2(i)
O2(i)–O2
3.316(5)
2.939(4)
3.314(5)
2.925(8)
3.314(5)
2.939(4)
3.316(5)
2.934(9)
O2–O5
3.060(6)
2.837(5)
3.053(6)
2.811(6)
2.811(6)
3.053(6)
2.837(5)
3.060(6)
O5–O7
refined unit-cell parameters for the HH phase are
3
˚
˚
˚
O7–O3
a ¼ 9:0168ð4Þ A, c ¼ 23:882ð1Þ A, V ¼ 1681:6 A (pattern
O3–O2
fitting reliability factor Rwp ¼ 13:6%). Those of the
O2(i)–O3(i)
O3(i)–O7(i)
O7(i)–O5(i)
O5(i)–O2(i)
˚
˚
LH phase are a ¼ 9:2311ð2Þ A, c ¼ 22:9296ð6Þ A, V ¼
3
˚
1692:2 A (reliability factor Rwp ¼ 8:9%). It can be noted
that the unit-cell parameters determined from the single-
crystal X-ray diffraction data are close to those of the HH
phase (Table 1). Moreover, the analysis of systematic
extinctions (00l: l ¼ 3n þ 1 in agreement with an helicoidal
axis along c) shows that the space group probably remains
unchanged. The parameter c is the most affected over the
Oxalate groups
C1–O1
1.271(5)
1.187(6)
1.263(5)
1.241(5)
1.579(8)
O1–C1–O2
129.5(3)
C1–O2
O3–C2–O4
126.0(5)
119.2(4)
114.8(4)
111.9(4)
121.1(4)
124.4(5)
126.2(6)
121.6(4)
112.0(4)
120.8(4)
114.7(4)
C2–O3
C1–C2–O3
C2–O4
C1–C2–O4
˚
C1–C2
C2–C1–O1
phase transition, with a decrease of ,1 A between the HH
C2–C1–O2
and LH phases. The a parameter varies in the opposite way
but with a lower magnitude. It can thus be concluded that
the overall crystal structure remains the same over the
transformation and that the rise of water content only causes
a change in the unit-cell parameters. An interesting feature
is, unlike what was observed for YK(C2O4)2·4H2O [27] or
Cd2Zr(C2O4)2·(4 þ x)H2O [1], the phase transformation is
here sudden. Indeed, in these two phases, the change in the
hydration level was followed by a proportional variation of
the unit-cell parameters. The LH phase can accommodate
up to a certain number of water molecules and when the
hydration reaches a critical level, it causes the structure to
switch to the HH phase.
C3–O5
1.237(5)
1.282(5)
1.197(5)
1.303(5)
1.576(8)
O5–C3–O6(iii)
O7–C4(iii)–O8(iii)
C3–C4(iii)–O7
C3–C4(iii)–O8(iii)
C4(iii)–C3–O5
C4(iii)–C3–O6(iii)
C3–O6(iii)
C4(iii)–O7
C4(iii)–O8(iii)
C3–C4(iii)
Symmetry codes: (i) 12y; 12x; 5=62z; (ii) x; x2y; 1-z; (iii) 1þx; y; z:
3.3. Subtle phase transformation
Powder diffraction data collected without paying atten-
tion towards ambient conditions (pressure and temperature)
revealed extra diffraction peaks than those expected from a
simulated powder pattern calculated from the structure
determined by single-crystal X-ray diffraction. These
supplementary diffraction peaks were successfully indexed,
with the program DICVOL91 [13], with an hexagonal
unit-cell characterized by parameters relatively close to
those of the phase previously described in this study
The thermal decomposition of CdZrCs1.5(H3O)0.5
(C2O4)4·x H2O was studied by thermogravimetric analysis
(Fig. 3) performed under air between 293 and 1173 K. The
humidity in air was ,70% during the recording and the
starting phase thus corresponds to the most hydrated one
(HH). The chemical formula derived for the HH phase is
CdZrCs1.5(H3O)0.5(C2O4)4·6.8H2O, which is consistent with
the formula derived from X-ray diffraction. Its decompo-
sition can be described by a four-step process. First, an
immediate weight loss is observed as soon as the product is
˚
˚
(a ¼ 9:029ð1Þ A; c ¼ 23:925ð9Þ A; Mð20Þ ¼ 21; Fð20Þ ¼
22ð0:010; 90Þ]. It was also noted that diffraction data
collections at different times showed a variation of

1218
E. Jeanneau et al. / Journal of Physics and Chemistry of Solids 65 (2004) 1213–1221
Fig. 2. Full-pattern matching of the X-ray powder diffraction diagrams obtained at 295 K for (a) the LH phase (37% humidity) and (b) the HH phase (65%
humidity). experimental pattern, calculated pattern and difference curve.
heated. This weight loss (,4.5%) corresponds to the
departure of about two water molecules. The TDXD plot
showed that this first weight loss corresponds to the
transformation of the HH phase into the LH one. A second
weight loss (,14%) is seen on the TG curve between ,333
and ,413 K, which is in agreement with the departure of
the remaining water molecules. The loss of these last water
molecules leads to an amorphisation of the compound, as
shown by the diffraction powder pattern. The third weight
loss observed between 443 and 643 K (,43%) is in good
agreement with the decomposition of the oxalate moieties
and the formation of cadmium oxide, zirconium oxide and
cesium carbonate (theoretical weight loss: 41.5%). An
X-ray powder diffraction pattern of the compound pre-
viously heated at 673 K confirmed the presence of ZrO2
[PDF2 No. 88-2390], CdO [PDF2 No. 78-0653] and Cs2CO3
[PDF2 No. 71-1981] [28]. The final weight loss (,69%)
should then correspond to the decomposition of the cesium
carbonate (theoretical weight loss: 71.6%). This thermal
analysis is in quite good agreement with the chemical
formula CdZrCs1.5(H3O)0.5(C2O4)4·x H2O proposed and,
furthermore, it allows to estimate the hydration difference
between the HH and LH phases, i.e. around two water
molecules.
the NMR properties of this nucleus more reminiscent of a
spin 1/2 than a quadrupolar spin 7/2 [33,34] and, usually, the
QI can be neglected.
Fig. 4a and b show the 133Cs static experimental spectra
obtained for the HH and LH phases, together with the
calculated spectra. The results for both phases are very close
to each other. They exhibit an original line shape where the
six external transitions, due to the first order QI, are well
resolved. In the middle appears the central transition
(1/2 ! 2 1/2) with a typical line shape due to the CSA.
Indeed, it is well known that QI, even with second order
effect, cannot give account of such a central line shape.
Finally, the whole static spectrum is perfectly reconstructed
by only taking into account the QI for the external transition,
on one hand, and the CSA for the central line on the other
hand. The parameters used for these reconstructions are
3.4. Solid-state 133Cs NMR
It is often assumed that the quadrupolar interaction
effects (QI) are predominant compared to chemical shift
anisotropy (CSA). However, it has been shown that in some
rare situations, CSA and QI have to be considered
simultaneously to take account of NMR spectra. Represen-
tative examples are 51V [29], 14N [30], 17O [31] and 133Cs
[32]. For 133Cs, the combination of a small quadrupole
moment (23 £ 10231 m2) and a large chemical shift range
(400 ppm due to the size of the electronic cloud) makes
Fig. 3. TG curve for CdZrCs1.5(H3O)0.5(C2O4)4·6.8H2O under flowing air
(heating rate: 5 K h21 between 293 and 673 K and 10 K h21 up to 1173 K).

E. Jeanneau et al. / Journal of Physics and Chemistry of Solids 65 (2004) 1213–1221
1219
Fig. 4. 133Cs NMR static spectra obtained for (a) HH and (b) LH phases together with their reconstruction.
gathered in Table 4. For each phase, LH or HH, an axial
anisotropy has to be considered simultaneously for QI and
CSA and the resultant isotropic chemical shifts for central
and external transitions are consistent, since they are very
close to each other (Table 4). These noteworthy 133Cs static
spectra should be put forward, since the CSA and the QI are
simultaneously visible and, moreover, independently influ-
ence the line shape of central and external transition lines.
To our knowledge, it is the first time that such a
phenomenon is observed on a NMR spectrum.
the tunnel along the [001] direction leads to a quick
exchange of the X and Y main axis of the QI and CS tensors
during the Csþ motion, whereas the Z-axis remains along
the [001] direction. Then, the NMR experiments give direct
access to the mean values of the chemical shift in the planes
perpendicular to [001], i.e. d’ ¼ ðdXX þ dYY Þ=2; and along
the [001] axis, i.e. d ¼ dZZ; leading to hc:s: ¼ 0: As VXX
¼
and VYY play an equivalent role, then hQ ¼ 0; and VZZ;
which provides the CQ value, coincides with the [001] axis
of the network. It can be noted that, recently, some authors
have reported similar effects for 129Xe in motion into
nanotubes [35,36].
In order to confirm this assumption, the 133Cs spectra of
the HH phase have been registered at various low
temperatures (Fig. 6). From room temperature to 130 K,
the line becomes broader and unresolved, like in amorphous
or glassy materials. This behavior is attributed to Csþ ions
freezing at random positions into the anionic tunnels,
leading thus to a distribution of the quadrupolar and
chemical shift parameters. Note that this behavior is
The MAS technique simultaneously averages the CSA
and the first order QI. The related spectra for the HH phase
(Fig. 5) exhibits a unique sharp line keeping the same
position whatever the spinning speed, proving that we
encounter only one mean type of cesium and confirming the
way the static spectrum was treated. For this line, we
measured the same isotropic chemical shift value than with
the static spectrum reconstruction (2200 ppm). An equiv-
alent result has been obtained for the LH phase. Note that
the LH isotropic chemical shift is about 15 ppm higher than
the one of the HH phase, whereas the CQ and dc:s:
parameters are stronger for the HH phase (Table 4).
At this stage, it is difficult to propose an explanation
concerning these values knowing that the sizes of the
tunnels containing the Csþ ions are almost the same.
Nevertheless, the perfect axial symmetry observed for both
QI electrical field gradient and CS is in agreement with the
motion of the Csþ ions into the one dimension channels
of the framework. Indeed, the helicoidal geometry of
Table 4
Reconstruction parameters for the 133Cs NMR spectra of the HH and LH
phases
External transition
Central transition
diso (ppm) CQ (kHz) hQ
diso (ppm) ds (ppm) hs
HH phase 2202
LH phase 2187
153
133
0.02 2199
2185
18
11
0.03
Fig. 5. 133Cs MAS NMR spectra of the HH phase at various spinning
speeds.
0
0

1220
E. Jeanneau et al. / Journal of Physics and Chemistry of Solids 65 (2004) 1213–1221
Finally, it is interesting to observe the 133Cs spectrum
recorded without paying any attention to temperature and
hygrometry (Fig. 7). The complex line shape obtained is
then the superposition of the two previous spectra attributed
to the pure LH and HH phases. The reconstructed line,
obtained by summing the components of the pure phases,
gives a good account of the experimental spectrum with
33% of HH and 66% of LH as relative intensities. An X-ray
diffraction pattern recorded with similar conditions con-
firmed the presence of these two phases, as already reported
above. The MAS spectrum recorded at a spinning rate of
3 kHz enables to confirm that, as expected, two types of
133Cs nucleus are present in the sample, with retrieved
isotropic chemical shifts corresponding to those found for
the pure phases.
Fig. 6. 133Cs NMR spectra of the HH phase collected at various
temperatures, from room temperature to 130 K.
3.5. 1H spectra
1
The H spectra of the powdered sample are given in
Fig. 8 together with their reconstruction. It is known that
the structure contains water molecules, as well as H3Oþ
cations to ensure the electro-neutrality of the material. The
coexistence of two chemical species containing 1H nucleus
in the structure is in agreement with the NMR obser-
vations. Indeed, to give account of the line shape, two
Lorentzian contributions, denoted a and b; are required.
They point out that protons have two types of behavior in
each crystalline phase. For both phases, the parameters
used for the reconstruction are very close to each other
(Table 5). Generally, the isotropic chemical shift of protons
in water is lower than in hydronium ions [37]. Conse-
quently, the a lines at 4 ppm are attributed to water and the
b lines at 9–10 ppm are assigned to H3Oþ. It appears that
the b lines are much broader than the a line, showing that
water molecules are in motion in the channel, whereas
H3Oþ ions are more bounded to the framework [37].
Following the same argument, it is clear that water
Fig. 7. 133Cs NMR spectra of CdZrCs1.5(H3O)0.5(C2O4)4·x H2O recorded
with a random hygrometric degree.
perfectly reversible and in full agreement with the motion of
Csþ ions in the anionic tunnels. When the sample returns to
room temperature, the spectrum recovers its initial specific
shape.
Fig. 8. 1H NMR spectra of (a) the HH and (b) the LH phases together with their reconstruction.

E. Jeanneau et al. / Journal of Physics and Chemistry of Solids 65 (2004) 1213–1221
1221
¨
[4] E. Jeanneau, N. Audebrand, M. Le Floch, B. Bureau, D. Louer, J. Solid
Table 5
Reconstruction parameters for the 1H NMR spectra of the HH and LH
State Chem. 170 (2003) 330–338.
[5] J.L. Hoard, J.V. Silverton, Inorg. Chem. 2 (1963) 235–243.
[6] C.W. Haigh, Polyhedron. 15 (1996) 605–643.
[7] Nonius, Kappa CCD Program Software, Nonius BV, Delft, The
Netherlands, 1998.
phases
diso (ppm)
Linewidth (kHz)
Intensity (%)
HH phase
LH phase
a line
b line
a line
b line
4
10
4
1.4
11.2
2.2
82
18
81
19
[8] Z. Otwinowski, W. Minor, Meth. Enzymol. 276 (1997) 307–326.
[9] P. Coppens, L. Leiserowitz, D. Rabinovich, Acta Crystallogr. 18
(1965) 1035–1038.
9
11.1
[10] L.J. Farrugia, J. Appl. Crystallogr. 32 (1999) 837–838.
[11] K. Brandenburg, M. Berndt, Diamond (version 2.1 e), Crystal Impact,
Bonn, 2001.
¨
[12] D. Louer, J.I. Langford, J. Appl. Crystallogr. 21 (1988) 430.
molecules are more mobile in the HH phase than in the LH
phase.
¨
[13] A. Boultif, D. Louer, J. Appl. Crystallogr. 128 (1991) 295.
[14] J. Rodriguez-Carvajal, Collected Abstracts of Powder Diffraction
Meeting, Toulouse, France, 1990, p. 127–128.
[15] T. Roisnel, J. Rodriguez-Carvajal, WinPLOTR: a windows tool for
powder diffraction, Mater. Sci. Forum 378–381 (2001) 118–123.
4. Conclusion
´ ´
[16] J. Plevert, J.P. Auffredic, M. Loue¨r, D. Loue¨r, J. Mater. Sci. 24 (1989)
1913–1918.
Compared to the compounds with a related hexagonal
structure with isotypic anionic [CdZr(C2O4)]22 frameworks
containing ammonium, amine or sodium, the cesium
compound reveals some peculiar features. Single-crystal
X-ray diffraction data together with 1H and 133Cs solid-state
NMR spectroscopy have shown that the cesium cations and
water molecules exhibits a dynamic disorder along the c-axis
of the hexagonal cell. Moreover, the cesium motion into the
channels with a hexagonal cross-section gives rise to a
noteworthy 133Cs NMR spectra, in which axial CSA and QI,
simultaneously and independently, influence theshapes of the
central and external transitions. Additionally, the diffraction
and NMR studies have clearly shown that the Cs compound
exhibits a phase transformation around ambient temperature,
which depends both on temperature and water vapor pressure.
It should be noted that the existence of a dynamic disorder
inside the channels is likely to be correlated with the ability
of the compound to exchange cations, from which the non-
stoichiometry of the material must originate.
[17] P.P. Man, J. Magn. Reson. 67 (1986) 78–85.
´
[18] D. Massiot, F. Fayon, M. Capron, I. King, S. Le Calve, B. Alonso, J.O.
Durand, B. Bujoli, Z. Gan, G. Hoatson, Magn. Reson. Chem. 40
(2002) 70.
[19] A. Altomare, M.C. Burla, M. Camalli, G. Cascarano, C. Giacovazzo,
A. Guagliardi, G.G. Moliterni, G. Polidori, R. Spagna, J. Appl.
Crystallogr. 32 (1999) 115–119.
[20] G.M. Sheldrick, SHELXL-97: Programs for Crystal Structure
¨
¨
Refinement, University of Gottingen, Gottingen, 1997.
[21] P.v.d. Sluis, A.L. Spek, Acta Crystallogr. A46 (1990) 194–201.
[22] Y. Lepage, J. Appl. Crystallogr. 20 (1987) 264–269.
[23] A.L. Spek, J. Appl. Crystallogr. 21 (1988) 578–579.
[24] C.W. Haigh, Polyhedron 14 (1995) 2871–2878.
[25] I.D. Brown, The chemical bond in inorganic chemistry. The bond
valence model, IUCr Monographs on Crystallography 12, Oxford
University press, New York, 2002.
[26] T. Hahn, Z. Anorg. Allgem. Chem. 109 (1957) 438–466.
´
¨
[27] T. Bataille, J.-P. Auffredic, D. Louer, Chem. Mater. 11 (1999)
1559–1567.
[28] Powder Diffraction File, International Centre for Diffraction Data,
Newtown Square, PA.
[29] U.G. Nielsen, J. Skibsted, H.J. Jakobsen, Chem. Phys. Lett. 356
(2002) 73–78.
[30] T. Giavani, H. Bildsoe, J. Skibsted, H.J. Jakobsen, J. Phys. Chem.
B106 (2002) 3026–3032.
Acknowledgements
[31] E. Scolan, C. Magnenet, D. Massiot, C. Sanchez, J. Mater. Chem. 9
(1999) 2467–2474.
The authors are indebted to Mr J.C. Ameline (GMCM,
University of Rennes), Dr T. Roisnel (CDIFX, University of
Rennes) and Mr G. Marsolier for their technical assistance.
[32] J. Skibsted, T. Vosegaard, H. Bildsoe, H.J. Jakobsen, J. Phys. Chem.
100 (1996) 14872–14881.
[33] H. Koller, B. Burger, A.M. Schneider, G. Engelhardt, J. Weitkamp,
Microporous Mater. 5 (1995) 219–232.
[34] D.K. Murray, J.W. Chang, J.F. Haw, J. Am. Chem. Soc. 115 (1993)
4732–4741.
References
[35] T. Meersmann, J.W. Logan, R. Simonutti, S. Caldarelli, A. Comotti,
P. Sozzani, L.G. Kaiser, A. Pines, J. Phys. Chem. A104 (2000)
11665–11670.
´
¨
[1] E. Jeanneau, N. Audebrand, J.P. Auffredic, D. Louer, J. Mater. Chem.
11 (2001) 2545–2552.
[36] P. Sozzani, A. Comotti, R. Simonutti, T. Meersmann, J.W. Logan, A.
Pines, Angew. Chem. Int. Ed. 39 (2000) 2695–2698.
´
¨
[2] E. Jeanneau, N. Audebrand, J.P. Auffredic, D. Louer, J. Mater. Chem.
12 (2002) 2283–2389.
¨
[3] E. Jeanneau, N. Audebrand, D. Louer, Chem. Mater. 14 (2002)
´ ´
[37] A. Abragam, Les principes du magnetisme nucleaire, Presses
Universitaires de France, Paris, 1961, p. 223.
1187–1194.