Analysis of the effects of thermal treatments on CaHPO4 by 31P NMR spectroscopy

Article abstract of DOI:10.1016/j.jallcom.2004.10.025

The compound CaHPO₄ was obtained by slow evaporation at room temperature. The sample was characterised by X-ray diffraction, differential scanning calorimetry and ³¹P MAS NMR spectroscopy at different annealing temperature. At room temperature, the observed values of the ³¹P NMR chemical shift for the title compound are found to be -1.6, -0.4 and 1.4 ppm with the proportions 1/4, 1/2 and 1/4, respectively, revealing the presence of three non-equivalent phosphorus sites in the structure. The ³¹P NMR investigation at different annealing temperatures points to a conversion of (HPO₄^2-) into (P₂O ₇^4-) at high temperature.

Full text of DOI:10.1016/j.jallcom.2004.10.025

Journal of Alloys and Compounds 394 (2005) 13–18
Analysis of the effects of thermal treatments on
CaHPO4 by 31P NMR spectroscopy
B. Louati, F. Hlel, K. Guidara∗, M. Gargouri
Laboratoire de l’e´tat Solide, Faculte´ des Sciences de Sfax, BP 802, 3018 Sfax, Tunisia
Received 16 September 2004; received in revised form 19 October 2004; accepted 19 October 2004
Available online 8 December 2004
Abstract
The compound CaHPO4 was obtained by slow evaporation at room temperature. The sample was characterised by X-ray diffraction,
differential scanning calorimetry and 31P MAS NMR spectroscopy at different annealing temperature. At room temperature, the observed
values of the 31P NMR chemical shift for the title compound are found to be −1.6, −0.4 and 1.4 ppm with the proportions 1/4, 1/2 and 1/4,
respectively, revealing the presence of three non-equivalent phosphorus sites in the structure. The 31P NMR investigation at different annealing
temperatures points to a conversion of (HPO42−) into (P2O74−) at high temperature.
© 2004 Elsevier B.V. All rights reserved.
Keywords: CaHPO4; Diphosphate; Thermal treatments; 31P NMR spectroscopy
1. Introduction
clear magnetic resonance) spectroscopic techniques. On the
one hand, at room temperature, the structural environments
of the P atoms have been analysed using NMR spectroscopy.
On the other hand, in order to investigate and discuss the ir-
reversible chemical process at high temperature consisting in
the formation of diphosphate Ca2P2O7, we have undertaken
CaHPO4 is of considerable biological importance in bones
and teeth, and finds practical uses in dental cements and
restorative materials [1]. More recently, it has been investi-
gated as protonic conductors [2,3]. The structures consist of
CaHPO4 chains bonded together by Ca · · · O bonds and three
types of hydrogen bonds [4,5]. Ca(1) is coordinated to seven
oxygen atoms in an approximately pentagonal bipyramid but
Ca(2) is coordinated to eight oxygen atoms. This sample ex-
hibits an irreversible chemical process at high temperature
consisting in the formation of diphosphate Ca2P2O7. The
complexmonophosphate(PO4)3− anddiphosphate(P2O7)4−
ions have been used as building blocks for a wide variety of
crystal phases with a wide spectrum of physical and chemical
properties. Both crystalline and glass monophosphate- and
diphosphate-based materials have gained much interest as
promising nonlinear optic materials, high-temperature ionic
conductors, solid electrolytes for high energy density batter-
ies, ion exchange materials and catalysts.
a
31P NMR study of a series of compounds with different
thermal treatments.
2. Experimental
2.1. Synthesis reaction
Alkaline earth hydrogen phosphate CaHPO4 was obtained
by spontaneous reaction at room temperature of a stoichio-
metric mixture of Ca(NO3)2 and NH4H2PO4 in water con-
form to the following scheme:
Ca(NO3)2 + NH4H2PO4
→ CaHPO4 + 2NO3− + NH4+ + H+
In this work, the CaHPO4 compound has been prepared
and characterised by XRD (X-ray diffraction) and NMR (nu-
Slow evaporation of the solution leads to the formation and
precipitation of the compound. The powder was warmed at
353 K for few hours to eliminate the humidity. The sample
∗
Corresponding author.
E-mail address: kamelguidara@yahoo.fr (K. Guidara).
0925-8388/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jallcom.2004.10.025

14
B. Louati et al. / Journal of Alloys and Compounds 394 (2005) 13–18
Table 1
Experimental conditions
ZG (MAS)
15.6
CP MAS
Pulse length (␮s)
Dead time (␮s)
Recycle time (s)
15.6
10
2
10
5
Resonance frequency (MHz)
MAS spinning speed (Hz)
Number of scan
Number of digitised points
Referencing 0 Hz
121.49
8000
720
4096
121.49
8000
32832
4096
H3PO4 (85%)
H3PO4 (85%)
pulse technique followed by 1H high-power decoupling. For
the recorded spectrum, a contact time of 1 ms and a period
between successive accumulations of 2 s were chosen. The
number of scans was 32832 (Table 1).
Fig. 1. Differential scanning calorimetry of CaHPO4 between 300 and
750 K.
was characterised by X-ray powder patterns, differential
scanning calorimetry and 31P NMR spectroscopy.
The structure of the sample consists of CaHPO4 chains
bonded together by Ca O bonds and three types of hy-
drogen bonds. Two distinct sets of pairs of PO4 units
are found in each primitive cell (Fig. 2). In the P(1)O4
group the O(2) P(1) O(4) angle is less than the tetrahe-
dral angle. Two other O P(1) O angles less than the tetra-
hedral angle involve oxygen O(1). In the P(2)O4 group,
P(2) O(6) and P(2) O(7) are the longest P O distances and
3. Results and discussion
3.1. Crystalline parameters
X-ray powder diffractogram reveals that the synthesised
compound is the anhydrous form of CaHPO4. It crystallises
˚
are about equal with an average value of 1.547 A. The only
O P O angle less than 109.5◦ in the P(2) O4 group in-
volves O(6) and O(7). Concerning the three types of hydro-
gen bonds, one type of hydrogen bond, O(1) H(2)· · ·O(5),
is normal but is at the short end of the normal range
¯
in the triclinic system (P1 space group) with unit cell param-
◦
˚
˚
˚
eters: a = 6.91(1) A, b = 6.◦63(3) A, c = 6.99(3) A, α = 96.32 ,
◦
3
˚
β = 103.87 and γ = 88.37 , V = 309.29 A which are in good
agreement with the literature values [2–4].
ꢀ
˚
with O(1)· · ·O(5) = 2.565(1) A, one, O(7)· · ·H(1)· · ·O(7 ),
ꢀ
˚
is very short with O(7)· · ·O(7 ) = 2.458(2) A, H(1) is
3.2. Calorimetric study
placed in the centre of symmetry in this hydro-
ꢀ
˚
gen bonds with O(7)· · ·H(1) = O(7 ) H(1) = 1.23 A and
The calorimetric study of CaHPO4 has been realised
between 300 and 750 K using a NETZSCH 204 differen-
tial scanning calorimeter. Fig. 1 represents the diagram ob-
tained at increasing temperature for a fresh preparation.
One endothermic region due to phase transformation in
the solid state of the sample was recorded. This region is
in the 550–750 K range with a large change of enthalpy,
ꢀH = 24 kJ/mol. The temperature anomaly is a combination
of two consecutive broad peaks at 642 and 678 K. The DSC
heating results are in agreement with the literature [6].
O(7) H(1) O(7ꢀ) = 180◦ [7] and one, O(6) H(3)· · ·O(8)
˚
where O(6)· · ·O(8) = 2.669(1) A, is the normal range but is
presumed to be statistically disordered with hydrogen cova-
lently bonded to half of the O(6) atoms on the average [4,5].
3.3. 31P NMR investigation
In order to study the temperature dependence of the
progress of the thermal decomposition, we have prepared a
set of samples with different thermal treatments. The species
are heated in air at atmospheric pressure for 24 h at fixed
temperatures ranging between 273 and 723 K. The 31P NMR
experiments are performed on a Bruker MSL 300 (B = 7.1 T)
spectrometer working at 121.49 MHz. The powdered sam-
ples are pocketed in the rotors and subjected to a spinning
speed of 8000 Hz. A ZG sequence program (zero go) are
used with a π/2 pulse length of 3 ␮s. NMR acquisition con-
ditions are reported in Table 1. The 31P CP MAS spectra
were obtained by means of the standard cross-polarisation
¯
Fig. 2. [0 1 0] view of the P1 structure of CaHPO4, emphasising the
hydrogen-bonding schema and the PO4 tetrahedra pattern.

B. Louati et al. / Journal of Alloys and Compounds 394 (2005) 13–18
15
Fig. 3. Deconvolution of the 31P NMR spectrum for the non-annealed com-
pound.
Fig. 4. 31P CP MAS NMR patterns of the non-annealed compound.
Fig. 3 represents the 31P NMR spectrum of the starting
The 31P NMR spectroscopy results seem to be in a dis-
agreement with the structural evidences. In fact, NMR in-
vestigation reveal the presence of three phosphorus sites but
only two distinct sets of pairs of PO4 units are found in the
primitive cell by X-ray investigation.
sample. The spectrum can be simulated by mixing three
Lorentzian functions at the positions −1.6, −0.4 and 1.4 ppm
in the proportions 1/4, 1/2 and 1/4, respectively. Simulation
data results are listed in Table 2, scatters and continuous line
correspond to the calculated and the experimental data, re-
spectively; the dotted lines represent the Lorentzian peaks.
It should be noted that the chemical shift of 31P MAS NMR
peaks corresponding to the non-annealed compound are in
agreement with the observations by several authors on differ-
ent apatites [8–10] and with the known hydroxyapatites (e.g.
Sr10(PO4)6(OH)2, Pb10(PO4)6(OH)2, etc.)whichδiso isinthe
range of 0.72; 2.9 ppm. The pure phase, (Cd10(PO4)6(OH)2),
is most deshielded of the hydroxyapatites and presents a
unique isotropic signal at δ = 12.2 ppm [11–13].
A careful study of the crystallographic data shows that
H(3) is presumed to be statistically disordered with hydrogen
covalently bonded to half of the O(6) atoms on the average.
Two configurations should, respectively, show H(3) bonded
to O(6) only and to O(6ꢀ) only. The two tetrahedra of P(2)
and P(2ꢀ) are inequivalent so when H(3) links to one of them,
two different groups HPO42− and H2PO4− are formed in the
2−
−
structure. In the NMR time scale, the HPO4 and H2PO4
entities are inequivalent and conduce to the appearance of
two distinct peaks in the spectrum, nevertheless the X-ray
diffraction picture is an average one, H(3) and H(3ꢀ) seem to
be simultaneously occupied in the X-ray time scale leading
to a single phosphorus type. In conclusion, in the NMR time
scale three distinct phosphorus types are present in the unit
crystallographic cell whereas two phosphorus types are de-
tectable by X-ray technique. It results that the B peak can be
unambiguously attributed to the P(1)O4 set of pairs and the
A and C peaks correspond to the P(2) and P(2ꢀ) tetrahedra.
At high temperature, the MHPO4 compounds
exhibit an irreversible process consisting in the
The 31P NMR spectrum of CaHPO4 reveals the pres-
ence of three distinct environments of (PO43−) in the
sample. In order to enhance the multinuclear signal in-
tensities, cross-polarisation (CP), involving the transfer
of magnetisation from abundant nuclei, usually from
protons, to the dilute nuclei (e.g. 31P, 13C, 29Si),
which also reduces the recycle delays, can be used.
The Hartmann–Hahn (HH) matching condition for CP,
ω(1H) = γ(1H)B1(1H) = γ(31P)B1(31P) = ω(31P), was opti-
mised for 31P NMR (Fig. 4). The CP NMR pattern is obvi-
ously characterised by the presence of three peaks placed at
−1.58, −0.3 and 1.37 ppm (Table 2). This result matches well
with the previously reported results and confirm the presence
of three phosphorus sites observed in the spectrum recorded
using (ZG) sequence program.
dehydration–condensation of the monohydrogen phos-
4−
phate leading to the formation of the diphosphate P2O7
.
This irreversible process occurs at T = 723 K in the case of
CaHPO4 [1,14].
Table 2
Deconvolution parameters of the 31P NMR spectrum for the starting compound
Peak
Position ( 0.5 ppm)
(ZG MAS)
Position ( 0.5 ppm)
(CP MAS)
Normalised area
(%) (ZG MAS)
Linewidth ( 10 Hz)
(ZG MAS)
A
B
C
−1.6
−0.4
1.4
−1.6
−0.3
1.4
26
47
27
176
342
240

16
B. Louati et al. / Journal of Alloys and Compounds 394 (2005) 13–18
Fig. 5. X-ray powder diffraction pattern of: (a) non-annealing compound CaHPO4 and (b) CaHPO4 annealed at T = 723 K for 24 h.
Table 3
Deconvolution parameters of the 31P NMR spectrum at different annealing temperatures for CaHPO4
T (K)
Line shape
Position ( 10 Hz)
Position ( 0.5 ppm)
Linewidth ( 10 Hz)
Normalised area (%)
300
Lorentz
Lorentz
Lorentz
−200
−50
170
−1.6
−0.4
1.4
176
342
240
26
47
27
373
473
573
Lorentz
Lorentz
Lorentz
−181
−26
136
−1.5
−0.2
1.1
182
247
362
43
30
27
Lorentz
Lorentz
Lorentz
−1043
−191
−37
−1317
−1081
−189
−68
−8.6
−1.6
−0.3
−10.9
−8.9
−1.5
−0.5
661
150
424
23
70
7
Lorentz
Lorentz
Lorentz
Lorentz
87
118
111
458
11
13
10
66
623
Lorentz
Lorentz
Lorentz
Lorentz
−1330
−1100
−189
−10.9
96
114
62
43
41
1
−9
−1.5
−1.1
−136
479
15
673
723
Lorentz
Lorentz
Lorentz
−1325
−1095
−418
−1325
−1096
86
−10.9
91
112
1537
43
43
14
−9
−3.4
Lorentz
Lorentz
Lorentz
−10.9
83
98
240
51
48
1
−9
0.7

B. Louati et al. / Journal of Alloys and Compounds 394 (2005) 13–18
17
Fig. 6. Deconvolution of the 31P NMR spectrum at various annealing temperatures. Inset: a zoom showing Q0 line.
In order to study the condensation of the phosphate in the
CaHPO4 compound, we have prepared a set of samples with
different thermal treatments. The species are heated in air at
atmospheric pressure for 24 h at fixed temperatures ranging
between 273 and 723 K. Fig. 5b shows the X-ray powder
pattern of the annealed compound at 723 K. All the peaks
can be indexed in the tetragonal system with the unit cell
of the ␤-calcium diphosphate (␤-Ca2P2O7) phase. The cell
parameters agree with the bibliographic data [15].
The set of the samples with different thermal treatment
is analysed by 31P NMR spectroscopy. Fig. 6 shows some
NMR spectra, clearly notable changes were recorded:
i. For T = 473 K, the B and C lines have merged into a sin-
gle line revealing the presence of drastically changes in
the P tetrahedra. This change is accompanied by the ap-
˚
˚
parameters (a = 6.68 A and c = 24.20 A, P41 space group).
These results show a single phase and confirm the formation

18
B. Louati et al. / Journal of Alloys and Compounds 394 (2005) 13–18
perature is plotted in Fig. 7. It increases with the annealing
temperature and presents an abrupt change in the slope of the
curve in the vicinity of 600 K. At T = 723 K, the reaction is
totally achieved and reaches the maximum value (98.8%).
4. Conclusion
In this work, we have synthesised the CaHPO4 compound.
The sample was investigated by X-ray diffraction, DSC and
31P NMR spectroscopy at different annealing temperatures.
31P MAS NMR study of the non-annealed compound
show three peaks A, B and C at the positions −1.6, −0.4
and 1.4 ppm in the proportions 1/4, 1/2 and 1/4, respectively,
revealing the presence of three types of phosphorus in the
structure. The B peak is ascribed to P(1)O4 group, however,
the A and C peaks correspond to the P(2) and P(2ꢀ) tetrahedra.
The DSC curve of CaHPO4 shows two endothermic peaks at
about 642 and 678 K. According to 31P NMR data evolu-
tion with annealing temperatures, the transition observed by
DSC has been interpreted as an irreversible chemical process
consisting in the formation of the diphosphate Ca2P2O7.
Fig. 7. Rate conversion of phosphate–diphosphate as a function of annealing
temperature.
pearance of a broad line denoted D at −8.6 ppm. The
large spectra modifications reflect that at high temper-
ature the CaHPO4 compound exhibits an irreversible
chemical process accompanied by a dramatic change in
the phosphorus environments due to the formation of
the diphosphate Ca2P2O7 conforming to the following
scheme:
2−
4−
2HPO4 → H2O + P2O7
References
ii. With increasing temperature, the intensity of the D line
increases at the cost of the A and C lines suggesting an
increase of the conversion rate. At high temperatures, it
splits into two components E and F of equal intensities
(Table 3). The E and F lines can obviously be attributed
to the phosphorus in different sites in the orthophos-
phate. The spectrum suggests the presence of a single
orthophosphate type in the previous structure. However,
the NMR results are insufficient to determine the phos-
phorus site that is the most sensitive to the conversion
process.
[1] G.R. Sivakumar, E.K. Girija, S.N. Kalkura, C. Subramanian, Cryst.
Res. Technol. 33 (1998) 197.
[2] L. Tortet, J.R. Gavarri, G. Nihoul, A.J. Dianoux, J. Solid State Chem.
132 (1997) 6.
[3] L. Tortet, J.R. Gavarri, G. Nihoul, A.J. Dianoux, J. Solid State Ionics
97 (1997) 253.
[4] B. Dickens, J.S. Bowen, W.E. Brown, Acta Cryst. B 28 (1971)
797.
[5] A. Boudjada, R. Masse, J.C. Guttel, Acta Cryst. B 34 (1978) 2692.
[6] R.A. Vargas, M. Mosquera, Revista Mexicana de Fisica 3 (1993)
450.
[7] M. Catti, G. Ferraris, Acta Cryst. B 33 (1977) 1223.
[8] J.L. Miquel, L. Facchini, A.P. Legrand, X. Marchandise, P. Lecouffe,
M. Chavanaz, M. Donozzan, C. Rey, J. Lemaˆıtre, Clin. Mater. 5
(1990) 111.
[9] J.P. Yenisowski, H. Eckert, J. Am. Chem. Soc. 109 (1987) 6274.
[10] A. Rodrigues, A. Lebugle, Colloid. Surf. A: Physicochem. Eng. As-
pects 145 (1998) 191.
[11] J.P. Yesinowski, J. Am. Chem. Soc. 103 (1981) 6266.
[12] B. Badraoui, M. Debbabi, R. Thouvenot, C.R. Hebd, Acad. Sci. Ser.
IIC 3 (2000) 107.
iii. The 31P MAS NMR spectrum recorded at 723 K shows
a smaller and narrow resonance at 0.7 ppm, Q0, and two
intense resonances E and F of equal intensities at −9
and −10.9 ppm, Q1. These resonances are assigned to
the orthophosphate and pyrophosphate species, respec-
tively. The chemical shifts seen for the last two reso-
nances are very close to the observed value of other acid
diphosphates [16,17]. In fact, the chemical shift values
of the Sm diphosphates SmHP2O7·3H2O(II) are −11.8
and −18.3 ppm and those of La HP2O7·3H2O(I) are −7.6
and −16.9 ppm [18].
[13] B. Badraoui, A. Bigi, M. Debbabi, M. Gazzano, N. Roveri, R. Thou-
venot, Eur. J. Inorg. Chem. 5 (2001) 1261.
[14] A. Durif, Crystal Chemistry of Condensed Phosphates, Plenum Press,
New York, 1995, pp. 23–57.
[15] N.C. Webb, Acta Crystallogr. 21 (1966) 942.
[16] N.J. Clayden, S. Esposito, P. Pernice, A. Aronne, J. Mater. Chem.
11 (2001) 936.
The percentage of diphosphate can be estimated from
NMR spectroscopy. It is represented by the ratio of the area
of the (E + F) peaks corresponding to the diphosphate to the
[17] P. Hartmann, J. Vogel, B. Schnabel, J. Non-Cryst. Solids 176 (1994)
157.
[18] S.B. Moussa, S. Ventemillas, A. Cabeza, E.G. Puebla, J. Sanz, Solid
State Chem. 177 (2004) 2129.
100(E + F)
total area of the spectrum, i.e.
. The rate
E + F + C + A + B
of conversion of phosphate–diphosphate as a function of tem-