Structural features of the reactives sites in α-M(DPO4)2 · D2O (M = Ti, Zr, Pb): Hydrogen-bond network and framework

Article abstract of DOI:10.1006/jssc.1996.0295

Three crystalline layered metal (IV) acid phosphates α-M(DPO₄)₂ · D₂O, α-MP, (M = Zr, Ti, Pb) have been hydrothermally synthesized and the size and shape of the microparticles have been analyzed by the Scherrer method. The crystal structures of α-ZrP and α-TiP have been refined from a combined X-ray and neutron powder diffraction study by the Rietveld method. These hydrogen phosphates are isomorphous and belong to the α-ZrP-type structure. However, this precise powder diffraction study has shown that the frameworks and hydrogen-bond networks of these materials are slightly different. The unit cell volume of α-ZrP is larger than that of α-TiP because the layers in α-ZrP are less corrugated than those in α-TiP. The H-bond network is also slightly different, as in α-ZrP there is not H bonding between the layers, the shortest contact being O(7) ... D(4) of 2.56(3) A. For α-TiP this contact is much shorter, 2.33(2) A, which may indicate a small but not negligible H-bond interaction between layers. Moreover, the intralayer H bonds are weaker in α-ZrP than in α-TiP as evidenced in the O(4) ... D(3) interactions of 2.41(3) and 2.16(2) A, respectively.

Full text of DOI:10.1006/jssc.1996.0295

JOURNAL OF SOLID STATE CHEMISTRY 125, 261–269 (1996)
ARTICLE NO. 0295
Structural Features of the Reactives Sites in ␣-M(DPO ) ؒ D O
4
2
2
(M ؍ Ti, Zr, Pb): Hydrogen-Bond Network and Framework
Enrique R. Losilla, Miguel A. G. Aranda, and Sebastian Bruque
Departamento de Qu ´ı mica Inorg a´ nica, Cristalograf ´ı a y Mineralog ´ı a, Universidad de M a´ laga, Aptd. 59, 29071 M a´ laga, Spain
Received April 1, 1996; in revised form June 13, 1996; accepted June 17, 1996
space and participate in the hydrogen-bond network (3,
Three crystalline layered metal (IV) acid phosphates ␣- 4). The hydrogen atom positions, although with large er-
M(DPO ؒ D O, ␣-MP, (M ؍ Zr, Ti, Pb) have been hydrother- rors, were derived from this powder neutron study. One
4
)
2
2
mally synthesized and the size and shape of the microparticles
have been analyzed by the Scherrer method. The crystal struc-
tures of ␣-ZrP and ␣-TiP have been refined from a combined
X-ray and neutron powder diffraction study by the Rietveld
method. These hydrogen phosphates are isomorphous and be-
long to the ␣-ZrP-type structure. However, this precise powder
diffraction study has shown that the frameworks and hydrogen-
bond networks of these materials are slightly different. The
unit cell volume of ␣-ZrP is larger than that of ␣-TiP because
the layers in ␣-ZrP are less corrugated than those in ␣-TiP.
hydrogen of the water acts as donor in one H bond with
the oxygen of a neighboring POH group. The other water
hydrogen does not participate in H bonding. Both crystal-
lographically independent acidic POH groups in the same
layer form hydrogen bonds with the oxygen of the water
molecule. The layers are only bound by van der Waals
forces. There are no neutron diffraction studies for other
members of the series. Therefore, it has been assumed,
so far, that the hydrogen bond network is similar in all
The H-bond network is also slightly different, as in ␣-ZrP there the series.
is not H bonding between the layers, the shortest contact being
We recently reported a preliminary structural study of
O(7) ؒ ؒ ؒ D(4) of 2.56(3) A. For ␣-TiP this contact is much Ͱ-MP, (M ϭ Ti, Sn, Pb, Zr) based on X-ray powder diffrac-
shorter, 2.33(2) A, which may indicate a small but not negligible tion (7). As a result of that study we concluded that the
˚
˚
H-bond interaction between layers. Moreover, the intralayer H
unit cell of Ͱ-ZrP is larger than those of the other members
of the series. However, some results were not conclusive
bonds are weaker in ␣-ZrP than in ␣-TiP as evidenced in
˚
the O(4) ؒ ؒ ؒ D(3) interactions of 2.41(3) and 2.16(2) A,
respectively.
as the derived parameters had large errors. It is important
to know the structural differences and similarities between
Ͱ-ZrP and the other members of the series and to under-
stand the H-bond network, as these materials are used as
acid solids (8), ion exchangers (9), sensors (10), and host
materials for intercalation reactions (11, 12).

1996 Academic Press, Inc.
INTRODUCTION
Acidic materials based on layered hydrogen phosphates
The structures are quite complex, with 16 atoms in the
asymmetric part of the unit cell, and we have therefore
carried out a combined powder X-ray and neutron diffrac-
tion study of Ͱ-ZrP and Ͱ-TiP to obtain a detailed picture
of the structures. This has allowed us to understand the
structural differences between Ͱ-ZrP and other members
of the series in the framework and the H-bond network.
of metals(IV) of composition M(HPO ) и H O (M ϭ Si,
4
2
2
Ge, Sn, Pb, Ti, Zr, Hf) constitute an isostructural series,
Ͱ-MP. Only Ͱ-ZrP has been structurally characterized by
single crystal X-ray diffraction (1, 2). Neutron powder dif-
fraction studies on Ͱ-ZrP have also been carried out (3,
4), however the precision on the reported structural data
were low. Other structural studies with powder data have
shown that for M ϭ Ti (5) and Hf (6) the frameworks are
similar to that of zirconium analog although the precision
of the reported data was quite low, probably due to the
low degree of crystallinity of the synthesized products.
The structure of Ͱ-ZrP is very well known (2) and is
EXPERIMENTAL
Synthesis
Deuterated powdered Ͱ-MP samples were synthesized
built up of layers of slightly distorted ZrO octahedra and under hydrothermal conditions. The starting compounds
6
alternating (up and down) HPO tetrahedra. These layers were D PO (85% w/w, 99ϩ% D, Aldrich-Chemie), D O
4
3
4
2
have a pseudo-hexagonal symmetry and are stacked along (99.8% D, Merck), and reactive anhydrous metal oxides
the c-axis. The water molecules are located in the interlayer synthesized as described below.
2
61
0022-4596/96 $18.00
Copyright  1996 by Academic Press, Inc.
All rights of reproduction in any form reserved.

262
LOSILLA, ARANDA, AND BRUQUE
˚
˚
Zr and Ti oxides were synthesized by slow hydrolysis b ϭ 5.294 A, c ϭ 15.447 A, and ͱ ϭ 101.69Њ for Ͱ-ZrP;
˚ ˚ ˚
of alcoholic metal alkoxide solutions (Zr(OCH CH CH )
3 4
a ϭ 8.640 A, b ϭ 5.009 A, c ϭ 15.510 A, and ͱ ϭ 101.32Њ
2
2
˚ ˚ ˚
0% and Ti[OCH(CH ) ] 97%). The white suspensions for Ͱ-TiP; and a ϭ 8.624 A, b ϭ 4.987 A, c ϭ 16.125 A,
3 2 4
7
were thoroughly washed with water and heated at 300ЊC and ͱ ϭ 100.61Њ for Ͱ-PbP.
for 2 h. X-ray powder patterns showed amorphous powders
and IR spectra indicated that these metal oxides were
RESULTS
anhydrous. PbO (Probus, analytical grade) was used as
2
source of lead, and it was heated to remove the ad-
sorbed water.
We have determined the average size of the microcrys-
tals of these materials from the X-ray powder data, and
Hydrothermal syntheses of Ͱ-MP (M ϭ Zr, Ti) were the crystal structures have been reanalyzed by the Rietveld
carried out in a Teflon-lined PARR bomb with a free method. GSAS allows refinement of a set of structural
volume of 45 ml. Zr(DPO ) и D O was obtained by heating parameters against different data sets, and in this case we
4
2
2
a mixture of 3 g of ZrO , 15 g of D PO , and 10 g of D O, have used neutron and X-ray powder diffraction data. The
2
3
4
2
at 165ЊC for 7 days. Ti(DPO ) и D O was obtained by previously determined structures of Ͱ-ZrP (2), Ͱ-TiP, and
4
2
2
heating a mixture of 2 g of TiO , 15 g of D PO , and 10 Ͱ-PbP (7) have been used as starting models for the re-
2
3
4
g of D O, at 150ЊC for 7 days. In both cases the overall finements. Initial hydrogen positions were taken from the
2
reactive molar ratios, M : P : D O, were 1 : 5 : 30.
previous neutron study (3) although they were confirmed
2
Pb(DPO ) и D O was hydrothermally synthesized in a by difference Fourier maps for each sample.
4
2
2
Teflon-lined BERGOFF reactor with a gas inlet and mag-
First, we refined the overall parameters for both data
netic stirring. The starting compounds were 2.51 g of PbO2 sets: background parameters, histogram scale factors, zero
and 50 g of D PO which result in an overall reactive molar shift errors, unit cell parameters, and peak shape parame-
3
4
ratio M : P : D O, of 1 : 40 : 40. Before heating, the air was ters for the pseudo-Voigt function. The refinements re-
2
removed and a pressure of 40 bar of O was introduced sulted in quite good fits. At this stage, sample dependent
2
in the reactor to avoid the reduction of the Pb(IV). Then, problems were attacked. These compounds are lamellar
the temperature was increased to 120ЊC for 6 days.
and the microparticles grow as plaques being longer along
The three resulting white solids were centrifuged and the a and b directions (parallels to the layers) than along
washed with 10 ml of D O, and two times with 30 ml of the c-axis (perpendicular to the layers), hence preferred
2
dried acetone. The solids were kept in a dessicator over orientation may be expected. We checked for this phenom-
KOH and a glass with 5 ml of D O. The X-ray powder enon and it was present only in the X-ray patterns. Neutron
2
diffraction patterns indicated the presence of highly crys- patterns did not present it as the samples were loaded
talline single phases.
into the cylindrical vanadium can without any pressing.
However, to get a flat surface for the X-ray experiment
the sample has to be pressed on the aluminium holder,
resulting in preferred orientation. GSAS can model this
Powder Diffraction
Neutron powder diffraction patterns were recorded on effect, through the March–Dollase correction (15), and
D1A diffractometer at ILL neutron source. The experi- the fit was slightly improved. Due to the shape of the
˚
mental conditions were: wavelength, 1.909 A; useful 2⌰ microparticles, described above, anisotropic peak broaden-
range 12Њ–140Њ, step size 0.05Њ (in 2⌰), counting for a day ing may also be expected. This phenomenon was observed
to record a high quality pattern. Three days were used to in both data sets, as the (hk0) peaks were slightly sharper
collect the patterns for Ͱ-MP (M ϭ Zr, Ti, Pb). X-ray than the (hkl) peaks with l ϶ 0. These can also be modeled
diffraction patterns for the same samples were recorded on with GSAS, resulting in slightly better fits. No absorption
a Siemens D501 automated diffractometer using graphite- correction was necessary for the neutron data.
˚
monochromated CuKͰ radiation, 1.5418 A. The data were
Second, the atomic parameters were refined: positional
collected in the Bragg–Brentano (⌰/2⌰) geometry (re- parameters, isotropic temperature factors, and D/H ratios.
flection mode). The experimental conditions were: 2⌰ We had to refine the deuterium/hydrogen ratios as the
range 15Њ–120Њ, step size 0.05Њ (in 2⌰), counting time 15 samples were not fully deuterated. To avoid correlations,
s, Ȃ9 h. The data were fitted by the Rietveld method (13) one isotropic temperature factor and one deuterium/hy-
using the GSAS suite of programs (14). The pseudo-Voigt drogen ratio was refined for each deuterium type (deute-
function corrected for asymmetry at low angles was used rium of the hydrogen phosphate group and of the water
to simulate the peak shape for both sets of patterns. The molecule).
background was fitted through a Fourier series by refining
five terms.
The average anisotropic microparticle sizes (for Ͱ-ZrP,
Ͱ-TiP, and Ͱ-PbP) were determined by the Scherrer
The X-ray patterns were indexed on monoclinic unit method (16) from the overall Lorentzian peak shape pa-
˚
cells, space group P2 /n, with dimensions: a ϭ 9.066 A, rameters obtained in the refinements of the X-ray patterns
1

STRUCTURE OF M(DPO
4
)
2
и D
2
O (M ϭ Ti, Zr, Pb)
263
TABLE 1
Peak Shape Parameters and Average Microparticle Sizes for
-MP (M ؍ Zr, Ti, Pb)
wider, due to a larger Gaussian contribution to the peak
shape. Hence the reported anistropic sizes of the micro-
crystals (Table 1) are those obtained from the X-ray pat-
terns.
␣
Zr
Ti
Pb
Hydrothermal syntheses led to microcrystals well devel-
a
2
oped (Ȃ900 A) along the ab plane. However, along the c-
˚
GW /Њ ϫ 10
37.8(6)
5.3(1)
8.9(6)
830(20)
310(15)
46.5(7)
5.0(1)
6.2(5)
880(20)
390(15)
89(4)
a
2
LX /Њ ϫ 10
STEC /Њ ϫ 10
4.8(4) axis direction (perpendicular to the layer) the size of the
59(3)
910(70)
70(5)
a
2
microcrystals were smaller. Although the temperature of
the synthesis for Ͱ-ZrP, 165ЊC, was higher than that for
Ͱ-TiP, 150ЊC, the size of the microcrystals along the c-axis
b
˚
P
P
ab /A Ȃ
b
˚
c
/A Ȃ
a
for Ͱ-ZrP was slightly smaller than that for Ͱ-TiP (see
Particle diameters along the crystallographic directions given. Calcu- Table 1). This is explained because the reactivity of the
Lorentzian peak shape parameters in GSAS (defined in Ref. 14).
b
lated from the Lorentzian peak shape parameters.
ZrO oxide precursor is smaller than that of the titanium
2
analog. It is worthy of mention that commercial crystalline
ZrO of small particle size (Alfa, 99ϩ, 1–3 Ȑm) does not
2
react under the hydrothermal conditions given in the Ex-
perimental.
To synthesize Ͱ-PbP, low temperature is necessary in
order to avoid partial reduction of Pb(IV) to Pb(II). This
low temperature, 120ЊC, results in a very small particle size
along the direction perpendicular to the layers, although
along the layer directions the crystals are well developed
(
14) and are given in Table 1. Final observed, calculated,
and difference profiles for Ͱ-ZrP and Ͱ-TiP are shown in
Figs. 1–4. Results of the combined powder diffraction study
are given in Table 2. Structural parameters are given in
Table 3 and selected framework bond distances and angles
are given in Table 4. Distances and angles around the water
molecule are given in Table 5. The observed, calculated,
and difference profiles for Ͱ-PbP are shown in Fig. 5.
(
Table 1). Under mild conditions, the crystals grow fast in
the ab plane and more temperature and time is necessary
to grow along the c-axis. However, both variables act
against the chemical stability of Ͱ-PbP.
DISCUSSION
The crystallinity and reactivity of these materials have
been a matter of much interest as they can be prepared in
Microparticles Study
The refinements of the neutron patterns gave similar a wide range of crystallinity depending upon the synthesis
refined values for the Lorentzian peak shape parameters conditions. The crystallinity (the microparticle sizes and
to those from the X-ray study. However, they were affected shapes) has important implications on the reactivity and
by a larger error because the neutron diffraction peaks are stability of these materials.
˚
FIG. 1. Final observed (points), calculated (full line), and difference neutrons (␭ ϭ 1.909 A) profiles for Ͱ-Zr(DPO
4
)
2
и D
2
O. Allowed reflection
marks are also shown.

264
LOSILLA, ARANDA, AND BRUQUE
˚
4 2 2
FIG. 2. Final observed (points), calculated (full line), and difference X-ray (␭ ϭ 1.5418 A) profiles for Ͱ-Zr(DPO ) и D O. Allowed reflection
marks are also shown.
Structural Study
tron data, and the contribution of heavy cations is stronger
in the X-ray patterns.
The refinement of a complex crystal structure against
In this study, we have used the same step size (0.05/
different data sets (X-ray and neutrons) obviously reduced 2⌰) for both patterns (X-ray and neutron) in order to
the systematic errors in powder diffraction studies, re- not overweight the X-ray data in the refinement. The
sulting in more accurate structural parameters. The com- decrease of the step size (adding more points to the
plementary information that can be extracted from both refinements) results in a decrease of the standard devia-
data sets is also important. Light element positions tions but this does not reflect increased accuracy of
(
hydrogen/deuterium) can be easily deduced from the neu- the refinement.
˚
FIG. 3. Final observed (points), calculated (full line), and difference neutrons (␭ ϭ 1.909 A) profiles for Ͱ-Ti(DPO
marks are also shown.
4 2 2
) и D O. Allowed reflection

STRUCTURE OF M(DPO
4
)
2
и D
2
O (M ϭ Ti, Zr, Pb)
265
˚
FIG. 4. Final observed (points), calculated (full line), and difference X-ray (␭ ϭ 1.5418 A) profiles for Ͱ-Ti(DPO
4
)
2
и D
2
O. Allowed reflection
marks are also shown.
The samples were highly deuterated. For Ͱ-TiP, the re-
finement converged to 83(2)% deuterium for the hydrogen
TABLE 3
Refined Atomic Parameters for ␣-TiP and ␣-ZrP
phosphate groups and 81(2)% deuterium for the water
molecule. For Ͱ-ZrP, these values were 88(4) and 96(4)%,
respectively. In this last case, a slight correlation between
the D/H ratios and their temperature factors (Table 1)
was detected, which increases the errors in both type of
parameters. However, this correlation does not affect the
˚
B(A )
2
Atom
M
x
y
z
Ti–NXPD
Zr–NXPD
Ti–NXPD
Zr–NXPD
0.24664(23)
0.24586(17)
0.3898(4)
0.2524(10) 0.48719(11) 0.42(4)
0.2537(9) 0.48535(8) 0.56(3)
0.7529(13) 0.39274(17) 0.69(7)
0.7509(22) 0.38589(23) 0.81(8)
0.2430(12) 0.40125(17) 0.53(6)
0.2457(22) 0.39756(23) 0.74(8)
P(1)
0.3882(5)
deuterium position and, hence, the H-bond network dis- P(2)
cussed below.
Ti–NXPD Ϫ0.13001(35)
Zr–NXPD Ϫ0.1325(4)
O(1)
Ti–NXPD
Zr–NXPD
Ti–NXPD
Zr–NXPD
Ti–NXPD
Zr–NXPD
Ti–NXPD
Zr–NXPD
0.5526(7)
0.5451(9)
0.3456(7)
0.3375(11)
0.2659(8)
0.2790(11)
0.3813(6)
0.3895(8)
0.8334(14) 0.4389(4)
0.8075(21) 0.4391(5)
0.4668(14) 0.4103(4)
0.4868(19) 0.4025(7)
0.9481(15) 0.4126(5)
0.9515(19) 0.4040(8)
0.94(16)
0.62(22)
0.64(14)
0.74(22)
1.11(16)
0.58(23)
O(2)
TABLE 2
O(3)
Some Results for the Combined X-Ray and Neutron Powder
Diffraction Study for ␣-MP (M ؍ Zr, Ti) in the Space Group
P 2 /n
1
O(4)
0.7583(20) 0.29069(31) 0.81(12)
0.7716(25) 0.2849(4)
0.4390(14) 0.4402(4)
0.4417(17) 0.4379(6)
Ϫ0.0431(14) 0.4368(4)
Ϫ0.0229(20) 0.4343(7)
0.59(17)
0.99(15)
0.34(22)
0.88(16)
1.66(27)
O(5)
O(6)
O(7)
O(8)
O(9)
D(1)
D(2)
D(3)
D(4)
Ti–NXPD Ϫ0.2309(8)
Zr–NXPD Ϫ0.2203(11)
Ti–NXPD Ϫ0.1439(8)
Zr–NXPD Ϫ0.1499(11)
Unit cell parameters
˚
˚
˚
a/A
b/A
c/A
ͱ/Њ
Ti–NXPD
Zr–NXPD
0.0408(7)
0.0334(8)
Ti–NXPD Ϫ0.1885(6)
Zr–NXPD Ϫ0.1943(8)
0.3384(13) 0.41553(32) 0.43(15)
0.3064(20) 0.4091(5) 0.27(22)
0.2488(20) 0.30063(31) 1.19(13)
0.2431(28) 0.2948(4)
0.7136(15) 0.2629(4)
0.7160(22) 0.2600(6)
Ͱ-ZrP refinement 9.0683(2)
5.2929(1)
5.0083(3)
15.4562(4)
15.4993(3)
101.727(1)
101.338(1)
Ͱ-TiP refinement
8.6358(2)
0.67(16)
1.13(13)
2.21(20)
3.04(34)
2.37(49)
3.04(Ϫ)
2.37(Ϫ)
3.58(42)
6.28(75)
3.58(Ϫ)
6.28(Ϫ)
Zr–N
pattern
Zr–RX
pattern
Ti–N
pattern
Ti–RX
pattern
Ti–NXPD
Zr–NXPD
Ti–NXPD
Zr–NXPD
0.0043(8)
0.0011(12)
0.4204(20)
0.4290(22)
0.917(4)
0.930(4)
0.2726(11)
0.2655(14)
˚
␭
/A
1.909
—
687
5.7
1.5418
0.960(2)
1066
1.909
—
637
5.4
1.5418
0.935(2)
975
Prefer orientation
No. hkl reflcn.
WP (%)
Ti–NXPD Ϫ0.2957(20)
Zr–NXPD Ϫ0.3044(20)
0.2363(35) 0.2783(11)
0.236(4)
0.874(4)
0.854(5)
0.703(4)
0.722(6)
0.2755(11)
0.2528(14)
0.2596(20)
0.3212(11)
0.3238(14)
R
R
R
R
7.8
7.5
Ti–NXPD
Zr–NXPD
Ti–NXPD
Zr–NXPD
0.0517(21)
0.0508(24)
0.0253(21)
0.0288(27)
P
(%)
(%)
2.5
4.3
5.6
3.1
3.0
2.1
7.2
4.2
2.5
3.1
F
overall
WP
(%)

266
LOSILLA, ARANDA, AND BRUQUE
TABLE 4
Framework Bond Distances and Angles for ␣-ZrP and ␣-TiP
Zr
Ti
Zr
Ti
M–O(1)
M–O(2)
M–O(3)
P(1)–O(1)
P(1)–O(2)
P(1)–O(3)
P(1)–O(4)
O(1)–M–O(2)
O(1)–M–O(3)
O(1)–M–O(5)
O(1)–M–O(6)
O(1)–M–O(7)
O(2)–M–O(3)
O(2)–M–O(5)
O(2)–M–O(6)
O(1)–P(1)–O(2)
O(1)–P(1)–O(3)
O(1)–P(1)–O(4)
O(2)–P(1)–O(3)
O(2)–P(1)–O(4)
O(3)–P(1)–O(4)
M–O(1)–P(1)
M–O(2)–P(1)
M–O(3)–P(1)
2.040(8)
2.070(11)
2.093(12)
1.521(9)
1.510(14)
1.516(12)
1.566(7)
90.3(4)
89.6(4)
89.7(4)
91.0(4)
178.6(6)
87.7(4)
1.929(6)
1.924(7)
1.940(8)
1.502(6)
1.521(9)
1.524(8)
1.569(5)
92.0(3)
90.4(3)
89.5(3)
89.1(3)
178.3(3)
88.2(3)
M–O(5)
M–O(6)
M–O(7)
P(2)–O(5)
P(2)–O(6)
P(2)–O(7)
P(2)–O(8)
O(2)–M–O(7)
O(3)–M–O(5)
O(3)–M–O(6)
O(3)–M–O(7)
O(5)–M–O(6)
O(5)–M–O(7)
O(6)–M–O(7)
2.043(10)
2.057(11)
2.062(7)
1.517(11)
1.551(14)
1.513(8)
1.574(7)
90.7(4)
177.3(5)
92.2(5)
89.5(4)
90.4(4)
91.2(4)
1.932(8)
1.920(8)
1.950(6)
1.514(8)
1.548(8)
1.525(7)
1.542(5)
89.4(3)
178.5(4)
92.3(4)
90.5(3)
89.2(3)
89.7(3)
89.8(5)
90.3(4)
88.0(4)
89.5(3)
178.7(4)
112.0(7)
109.7(8)
109.2(5)
112.6(7)
107.4(8)
105.8(8)
159.4(9)
146.3(7)
145.5(7)
178.8(4)
114.3(5)
111.1(5)
109.2(4)
110.9(5)
103.5(5)
107.3(5)
150.6(5)
142.6(5)
140.4(5)
O(5)–P(2)–O(6)
O(5)–P(2)–O(7)
O(5)–P(2)–O(8)
O(6)–P(2)–O(7)
O(6)–P(2)–O(8)
O(7)–P(2)–O(8)
M–O(5)–P(2)
112.1(7)
113.5(8)
108.2(8)
108.9(7)
108.6(9)
105.2(5)
155.5(7)
148.5(7)
146.1(6)
111.5(5)
110.8(5)
106.9(5)
112.3(5)
110.1(6)
104.8(4)
149.6(5)
143.2(5)
139.6(4)
M–O(6)–P(2)
M–O(7)–P(2)
Unfortunately, there was a problem in the synthesis of
Framework of Ͱ-MP series. The structural results for
the lead hydrogen phosphate that resulted in a poorly Ͱ-ZrP are essentially identical to those previously reported
deuterated sample. Due to the low degree of deuteration from single crystal data (2). All framework positional pa-
and to the small particle size along the c-axis, the neutron rameters obtained in this study (Table 3) agree, within 3–4
data was of insufficent quality to derive good structural standard deviations, with the single crystal values. The
parameters. As the aim of this work is to analyze precisely refinement for Ͱ-TiP converged to a very similar structure.
the crystal structures of these phosphates, the combined MO octahedra and PO tetrahedra are regular as deduced
6
4
neutron and X-ray study for the lead compound will not from the distances and angles given in Table 4. It is worth
be reported.
pointing out that P–O bond distances to the oxygen of the
TABLE 5
Distances and Angles Around the Water Molecules for ␣-ZrP and ␣-TiP
Zr–NXPD
Zr–NPD
Ti–NXPD
Zr–NXPD
Zr–NPD
Ti–NXPD
O(4)–D(1)
0.982(20)
0.860(26)
1.718(23)
2.408(31)
175(2)
159(3)
122(2)
117(2)
0.91(7)
0.94(7)
1.92(7)
2.16(7)
169(7)
160(7)
—
0.930(18)
0.927(20)
1.751(20)
2.158(22)
179(2)
168(2)
118(1)
109(1)
O(8)–D(2)
O(9)–D(4)
O(9) и и и D(2)
0.983(17)
0.968(23)
1.759(20)
2.561(32)
179(2)
112(1)
103(2)
87(3)
1.05(6)
0.92(5)
1.78(6)
—
159(6)
—
0.925(16)
0.887(17)
1.792(17)
2.325(19)
179(2)
111(1)
105(2)
103(2)
O(9)–D(3)
O(9) и и и D(1)a
a
O(4) и и и D(3)a
O(7) и и и D(4)b
O(4)–D(1) и и и O(9)
O(9)–D(3) и и и O(4)
D(1)–O(9) и и и D(3)
D(2)–O(9) и и и D(3)
O(8)–D(2) и и и O(9)
D(1)–O(9)–D(2)
D(1)–O(9) и и и D(4)
D(3) и и и O(9) и и и D(4)
—
—
—
Note. The values reported from the powder neutron diffraction study, Zr-NPD (3), are given for comparison.
a
Intralayer H bonds.
Interlayer H bonds.
b

STRUCTURE OF M(DPO
4
)
2
и D
2
O (M ϭ Ti, Zr, Pb)
267
˚
FIG. 5. Final observed (points), calculated (full line), and difference X-ray (␭ ϭ 1.5418 A) profiles for Ͱ-Pb(DPO
4
)
2
и D
2
O. Allowed reflection
marks are also shown.
hydrogen phosphate groups are slightly longer than those (3) in the intralayer H bond between D(3) and O(4) in Ͱ-
˚
to the remaining oxygens, for both Ͱ-ZrP and Ͱ-TiP, as ex- ZrP. The refined distance is 2.41(3) A, previously reported
˚
xpected.
as 2.16(7) A. This intralayer H bond is stronger for a-TiP
˚
One of the objectives of this study was to compare the as evidenced in the refined distance 2.16(2) A, and it is
frameworks of Ͱ-ZrP and Ͱ-TiP, because in our previous more linear.
work (7) we concluded that Ͱ-ZrP has a larger unit cell
In Fig. 6, the location of the water molecule between
than other members of the series. The comparison of the two Ͱ-MP layers can be seen. The very strong and linear
framework bond distances and angles given in Table 4 H bonds, O(9)–D(1) and O(9)–D(2), are also displayed.
clearly shows that the M–O–P bonds are more open for The D(4) of the water molecule points to the center of
Ͱ-ZrP than for Ͱ-TiP as previously suggested in our X-ray the zeolitic cavity defined by the three hydrogen phosphate
study of Ͱ-TiP, Ͱ-PbP, and Ͱ-SnP (7). Hence, it can be groups but it does not interact strongly with that layer.
concluded that the Ͱ-ZrP layers are less corrugated than The H-bond network along the c-axis is slightly different
those of the other members of the series. In other words, in these materials. It has been confirmed that there are no
the phosphate tetrahedra are more extruded from the lay- H bonds between the layers in Ͱ-ZrP. The shortest contact
˚
ers for Ͱ-MP (M ϭ Ti, Pb, Sn) than for Ͱ-ZrP (and Hf). is O(7) и и и D(4), 2.56(3) A, which is too long to be consid-
These flat layers in Ͱ-ZrP result in more free area per ered as a hydrogen bond. Hence, the layers in Ͱ-ZrP are
P–OH group. As these groups are the active center in only held together by van der Waals interactions. However,
˚
these materials, the high reactivity of Ͱ-ZrP is structur- for Ͱ-TiP this contact is much shorter, 2.33(2) A, which
ally justified.
may indicate a small but not negligible H-bond interaction
between layers. The results are in full agreement with our
early report (7) of a stronger retained water molecule in
Ͱ-TiP than that in Ͱ-ZrP, derived from an IR and thermal
study. In the previous study (7) we concluded that Ͱ-ZrP
was the member of the studied series with the least retained
H-bond network in Ͱ-MP series. The deuterium posi-
tions derived from this neutron study are similar to those
obtained in the previous neutron work (3, 4). However,
the better quality of our patterns and the possibility to
compare Ͱ-ZrP and Ͱ-TiP has allowed us to understand
better the H-bond network in these materials. First, we hydration water.
have confirmed the two more important H bonds that take
The deuterium environment around the water molecule
place between the deuteriums of the hydrogen phosphate is a distorted tetrahedron with two short bonds, to D(3)
groups, D(1) and D(2), and the oxygen of the water mole- and D(4), and two very long bonds to D(1) and D(2). The
cule, O(9). These two hydrogen bonds are in the same angles are quite close to the expected tetrahedral value,
layer and are linear (Table 5, see Fig. 6). However, our 109Њ, and are more regular for Ͱ-TiP than for Ͱ-ZrP
study does not fully agree with the previous neutron work (Table 5).

268
LOSILLA, ARANDA, AND BRUQUE
FIG. 6. Location of the water molecule in the interlayer region of Ͱ-MP, showing the H-bond network. The linear H bonds, O(9)–D(1) and
O(9)–D(2), are displayed as narrow solid lines.
There are three main structural characteristics that make
Ͱ-ZrP a singular, more reactive, and stable host for interca-
lation reactions. (i) It has more free area per P–OH group,
allowing large species to be intercalated. (ii) The layers
are flatter improving the accesibility of the guest com-
ACKNOWLEDGMENTS
We thank Dr. Alan Hewat (ILL) for his help during the neutron
diffraction experiment (5-22-413), and to ILL for the provision of neutron
facilities. This work was supported by DGICYT (Ministerio de Educaci o´ n
y Ciencia, Spain) research grant PB93/1245. E. R. L. thanks the Junta
pounds and minimizing the steric hindrance of the guest– de Andaluc ´ı a (Spain) for a studentship.
host complex. (iii) The layers are weaker bonded making
possible facile intercalation reactions under the same con-
ditions and with similar microparticles sizes. All these
structural features make the intercalation chemistry of Ͱ-
REFERENCES
ZrP richer and relatively easier.
1
2
. A. Clearfield and G. D.Smith, Inorg. Chem. 8, 431 (1969).
. J. M. Troup and A. Clearfield, Inorg. Chem. 16, 3311 (1977).
CONCLUSIONS
3. J. Albertsson, A˚ . Oskarsson, R. Tellgren, and J. O.Thomas, J. Phys.
Chem. 81, 1774 (1977).
Although acid layered Ͱ-MP materials have similar crys-
tal structure, some structural differences have been found
through combined X-ray and neutron powder diffraction
study. There are two main structural differences between
Ͱ-ZrP and other members of the series. First, the layers
in Ͱ-ZrP are more flat resulting in more free area per
P–OH group. Second, the H-bond network are slightly
different. There are no H bonds between the layers, in Ͱ-
ZrP, which are therefore held together only by van der
Waals forces. However in Ͱ-TiP, and probably in other
members of the series, there is a small but not negligible
H-bond interaction between layers.
4. A. Christensen, E. K. Andersen, I. G. K. Andersen, G. Alberti, M.
Nielsen, and M. S. Lehmann, Acta Chem. Scand. 44, 865 (1990).
5
. I. Nakai, K. Imai, T. Kawashima, K. Ohsumi, F. Izumi, and I. Tomita,
Anal. Sci. 6, 689 (1990).
6
7
. G. Schuck et al. Solid State Ionics 77, 55 (1995).
. S. Bruque, M. A. G. Aranda, E. R. Losilla, P. Olivera-Pastor, and
P. Maireles-Torres, Inorg. Chem. 34, 893 (1995).
8. T. Hattori, A. Ishiguro and Y. Murakami, J. Inorg. Nucl. Chem. 40,
107 (1978).
1
9
. A. Clearfield, in ‘‘Inorganic Ion Exchange Materials.’’ CRC Press,
Boca Raton, FL, 1982.
1
0. G. Alberti, F. Cherubini, and R. Palombari, Sensors Actuators B
4–25, 270 (1995).
11. G. Alberti, in ‘‘Recent Developments in Ion Exchange’’ (P. A.
2

STRUCTURE OF M(DPO
4
)
2
и D
2
O (M ϭ Ti, Zr, Pb)
269
Williams and M. J. Hudson, Eds.). Elsevier Applied Science, Amster-
dam, 1987.
14. A. C. Larson and R. B. Von Dreele, Los Alamos National Laboratory
Report No. LA-UR-86-748, 1987.
1
1
2. E. Rodriguez-Castellon, A. Rodriguez, and S. Bruque, Inorg. Chem.
15. (a) W. A. Dollase, J. Appl. Cryst. 19, 267 (1987); (b) A. March, Z.
Krystallogr. 81, 285 (1932).
16. P. Scherrer, Nachr. Ges. Wiss. Gottingen, 98 (1918).
2
4, 1187 (1984).
3. H. M. Rietveld, J. Appl. Cryatallogr. 2, 65 (1969).