JOURNAL OF SOLID STATE CHEMISTRY 135, 96—102 (1998)

ARTICLE NO. SC977602

XRD and Solid-State NMR Study of Magnesium Oxide–Magnesium

Orthophosphate Systems
Maria Angeles Aramendı́a, Victoriano Borau, César Jiménez,1 José Maria Marinas,
Francisco José Romero, and José Rafael Ruiz
Department of Organic Chemistry, Faculty of Sciences, Co& rdoba University, Av. San Alberto Magno s/n, E-14004 Co& rdoba, Spain

Received March 7, 1997; in revised form August 15, 1997; accepted August 19, 1997

research, authors have used a variety of physical techniques

Various magnesium oxide–magnesium orthophosphate sys- including IR, visible, UV, and Raman spectroscopies, in
tems in different Mg/P ratios were synthesized and examined by addition to ESR, ESCA, EXAFS, and XRD, among others,
XRD and 1H and 31P MAS NMR. The results thus obtained are for this purpose. Specifically, 1H MAS NMR spectroscopy
compared with those for commercially available pure MgO and was recently used for the structural elucidation of Al O
Mg3(PO4)2. The crystallization temperature of magnesium or- 2 3
(13), TiO —SiO (14), and V O —AlPO systems (15). Also,
thophosphate was found to increase with increasing MgO con- 2 2 2 5 4
31P MAS NMR has enabled the characterization of the
tent in the sample. The NMR technique revealed the absence of
local environment of phosphorus atoms in both amorphous
interactions between the two species. 31P MAS NMR allowed the
amorphous-to-crystalline ratio for magnesium orthophosphate in
and crystalline phosphates (16—19).
solids calcined at different temperatures to be determined. Also, This work reports on the synthesis and characterization
1
H MAS NMR revealed the qualitative nature of hydroxyl sites of magnesium oxide—magnesium orthophosphate systems
in the solids. ( 1998 Academic Press by use of various instrumental techniques. The results thus
obtained are compared with those for commercially avail-
able magnesium oxide and magnesium orthophosphate.
INTRODUCTION

EXPERIMENTAL
The literature on the use of metal orthophosphates as
catalysts for organic processes grows day by day. Thus, Materials
AlPO (1—4), BiPO (5—7), BPO (8), FePO (9), Li PO
4 4 4 4 3 4 The magnesium oxide—magnesium orthophosphate sys-
(10), and Mg (PO ) (11) have been used in organic syn-
3 42 tems studied in this work are designated Mg P , where
theses involving catalysis by acid or acid and basic sites. 100~x x
Magnesium orthophosphates have not yet been fully ex- x is the percent mole fraction of magnesium phosphate and
plored as catalysts for organic processes, however. Our 100!x that of magnesium oxide in the solid. Thus, the
solid named Mg P was obtained by dropwise addition of
group has used various magnesium orthophosphates ob- 50 50
a solution containing 8.6 ml of H PO (o"1.8 g/cm3) in
tained by modified synthetic procedures in the gas-phase 3 4
dehydration—dehydrogenation of cyclohexanol (12). 50 ml of distilled water over a suspension of 22.2 g of
Mg(OH) in 50 ml of distilled water in an ice bath under
In recent years, solid-state NMR spectroscopy has grown 2
enormously in use for characterizing solids, in combination continuous stirring that was extended for a further 1 h after
with diffraction techniques for examining crystalline phos- addition was finished. The resulting suspension was vac-
uum-dried. Solids Mg P and Mg P were obtained by
phates. Also, solid-state NMR spectroscopy has become 25 75 75 25
the same procedure but using 22.2 g of Mg(OH) and
a crucial tool for identifying amorphous compounds, which 2
12.9 ml of H PO for the former and 22.2 g of Mg(OH) plus
is unaffordable by X-ray diffraction techniques, as it allows 3 4 2
4.3 ml of H PO for the latter.
the local environment of atoms with a potential catalytic 3 4
activity to be characterized. Solids MgP (Aldrich 34,470-2) and MgO (Probus 3225)
were commercially available products of formula Mg
Because characterizing the chemical nature and proper- 3
(PO ) ) xH O and Mg(OH) , respectively.
ties of surface active sites is one of the main goals of catalysis 42 2 2
All solids were sieved through 200—250 mesh and dried in
a stove at 293 K prior to calcination at 623, 773, 923, or
1 To whom correspondence should be addressed. 1073 K.
96
0022-4596/98 $25.00
Copyright ( 1998 by Academic Press
All rights of reproduction in any form reserved.
 MgO—Mg (PO ) SYSTEMS
3 42
97

X-Ray Diffraction Analysis

X-ray diffraction (XRD) patterns were recorded on a
Siemens D500 diffractometer using CuK radiation. Scans
a
were performed over the 2h range from 7° to 50°.

NMR Spectra
Solid-state 1H and 31P NMR spectra were recorded
on a Bruker ACP-400 spectrometer at 400.13 and
161.975 MHz, respectively, at room temperature. Typical
spinning speeds (3.5 kHz) were used with all samples. The
excitation pulse and recycle time for 1H NMR were 5 ls (n/2
pulse) and 1 s (1000 scans), respectively—longer delay times
had little effect on the relative intensities of proton reson-
ances. The excitation pulse and recycle time for 31P NMR
were 5 ls (n/2 pulse) and 3 s (300 scans), respectively. The
contact time for the transfer of magnetization between pro-
tons and 31P (CP experiments) was 10 ms (300 scans).
1H and 31P shifts were measured relative to tetramethyl-
silane and H PO , respectively. All samples were degassed
3 4
to 3 lm Hg at 393 K for 8 h before their spectra were re-
corded.

RESULTS AND DISCUSSION

XRD Results FIG. 1. XRD patterns for solid MgP, uncalcined (a) and calcined at 393
(b), 623 (c), 773 (d), 923 (e), and 1073 K (f ).
Figure 1 shows the XRD patterns for solid MgP heated at
different calcination temperatures. The uncalcined solid ex-
hibited bands consistent with the formula Mg (PO ) )
3 42
8H O. As can be seen in the figure, the solid was amorphous MgO and farringtonite Mg (PO ) . Kanazawa et al. (20)
2 3 42
at 393 and 623 K and exhibited some crystallinity above reported the transition from Mg (PO ) ) 22H O to
3 42 2
773 K. The crystalline structure was identified as Mg Mg (PO ) ) 8H O to occur at 328 K in a water-rich atmo-
3 3 42 2
(PO ) (farringtonite phase) over the temperature range sphere. We could not identify such a transition or the
42
923—1073 K. phosphate octahydrate on raising the calcination temper-
Figure 2 shows the patterns corresponding to solid MgO ature, although we cannot rule out its occurrence. It is
calcined at different temperatures. The uncalcined solid logical to think that the octahydrate (whether crystalline or
consisted of brucite Mg(OH) . On calcination at 623 K it otherwise) may have been formed as the calcination temper-
2
exhibited primarily Mg(OH) bands and incipient periclase ature was raised in our case before the anhydrous magne-
2
MgO bands. At 773 K, the solid consisted virtually com- sium orthophospate reached by subjecting the solids to high
pletely of periclase; also, brucite bands were almost absent calcination temperatures appeared.
from the patterns. Finally, at 923 and 1073 K, the solid The presence of MgO in different proportions in the solid
consisted of highly crystalline periclase MgO. had an appreciable effect on the thermal transformation of
Figure 3 shows the XRD patterns for magnesium ox- magnesium orthophosphate. The formation of interstitial
ide—magnesium orthophosphate in a Mg P sample. All MgO inhibits or delays—depending on the particular
50 50
uncalcined solids exhibited the diffraction bands for brucite solid—the crystallization of Mg (PO ) species.
3 42
Mg(OH) and Mg (PO ) ) 22H O. Calcination led to Uncalcined solids can be assumed to contain the species
2 3 42 2
amorphization of all the solids; the band for periclase MgO brucite Mg(OH) and Mg (PO ) ) 22H O. When these
2 3 42 2
was clearly observed above 773 K. The higher the propor- solids are calcined at 773 K, the sole identifiable bands are
tion of magnesium oxide in the solid was, the stronger were those for periclase MgO on an amorphous background—
these bands. Magnesium orthophosphate was amorphous in fact, magnesium orthophosphate is amorphous at
in all solids except those treated at 923 K, where it appeared that temperature. The orthophosphate bands were also
to be slightly crystalline. The pattern in Fig. 3c reveals that hardly distinguishable in the patterns obtained on calcina-
solid Mg P calcined at 923 K was a mixture of periclase tion at 923 K. It was shown elsewhere that magnesium
50 50
98 ARAMENDÏA ET AL.

FIG. 2. XRD patterns for solid MgO, uncalcined (a) and calcined at
623 (b), 773 (c), 923 (d), and 1073 K (e).

orthophosphate changes from amorphous to crystalline be-

low 923 K; consequently, the presence of MgO in these
systems may delay the crystallization of amorphous magne-
sium orthophosphate.

31
FIG. 4. 31P NMR spectra for solid MgP, uncalcined and calcined at
P NMR Spectroscopy different temperatures: (a) MAS spectra; (b) CP/MAS spectra. *Room
temperature.
Figure 4a shows the 31P MAS NMR spectra for solid
MgP calcined at different temperatures. As can be seen, the
spectra consist of a single component and a symmetric series
of spinning sidebands (ssb). The behavior of this signal at
different temperatures was consistent with previous XRD
findings. Thus, the chemical shift and width for this
component changed with the thermal treatment applied
(Table 1) as a result of the conversion from a crystalline

TABLE 1
Chemical Shifts (in ppm Relative to 85% w/w H3PO4) and
Widths of the Signals for Solid MgP Calcined at Different
Temperatures

¹ (K) d 31P (ppm) l (Hz) Crystal phase

Room 4.6 732 Mg (PO ) ) 8H O

3 42 2
623 0.5 1463 Amorphous
773 0.5 1458 Amorphous
!0.5 134 Farringtonite
923 !0.5 132 Farringtonite
FIG. 3. XRD patterns for solid Mg P , uncalcined (a) and calcined
50 50 1073 !0.5 130 Farringtonite
at 773 (b) and 923 K (c).
 MgO—Mg (PO ) SYSTEMS
3 42
99

that the experimental line can be decomposed into two

Gaussian lines (Fig. 5) with different line widths and chem-
ical shifts (Fig. 4c), indicating the presence in this sample of
two different types of 31P nuclei, corresponding to the
amorphous (0.5 ppm) and crystalline (farringtonite) magne-
sium orthophosphate (!0.5 ppm). This Gaussian decon-
volution allowed the width, chemical shift, and area for each
signal in this spectrum to be determined; in this way, the
solid was found to be 76% amorphous and 24% crystalline.
The solids calcined at 923 and 1073 K gave a sharp signal
corresponding to farringtonite magnesium orthophosphate,
consistent with a previous finding from the XRD patterns
(Figs. 1d and 1e), which exhibited weak signals that
FIG. 5. 31P MAS NMR spectrum of MgP calcined at 773 K together strengthened as the calcination temperature was increased
with its decomposition into Gaussian lines. (to 923 K). These results suggest that the crystalline and
amorphous phases of magnesium orthophosphate already
coexist at 773 K. This was previously inferred from a SEM
phase to an amorphous phase and back to a new crystalline study (12) that revealed the presence of two types of par-
phase. The spectrum for the uncalcined sample exhibited ticles: flat, oblong or in the shape of elongated rods.
a chemical shift of 4.6 ppm that dropped to 0.5 ppm on To identify the phosphorus atoms bonded to OH groups,
calcination at 623 K. The spectrum for the sample calcined the 31P CP/MAS spectra shown in Fig. 4b were recorded. In
at 773 K exhibited two overlapped signals. It can be seen this technique, the polarization transfer from 1H nuclei to

FIG. 6. 31P MAS spectra for solid Mg P , uncalcined (a) and calcined at 623 (b), 773 (c), 923 (d), and 1073 K (e).
100~x x
100 ARAMENDÏA ET AL.

31P nuclei takes place by irradiation of the two types of

nuclei at their respective resonance frequencies under the
Hartmann—Hahn condition (21, 22). Spinning the sample
and irradiating the protons while the 31P signal is acquired
reduce 1H—31P dipole interactions and sharpens the phos- (c)
phorus signal. The uncalcined sample exhibited a sharp
signal (163 vs 732 Hz in the MAS spectrum) of greater
height, which suggests the presence of OH groups bonded to (b)
phosphorus throughout the crystal lattice. As the calcina-
tion temperature was raised, the phosphorus signal de-
creased and eventually disappeared at 923 K by the effect of (a)
the sample dehydroxylation. The 31P CP/MAS spectrum
obtained at 923 K contained a single signal (0.5 ppm), thus
indicating that it was the phosphorus in the amorphous
phase that bonded to OH groups.
These results are quite consistent with those previously FIG. 7. 1H MAS NMR spectra for solid MgO, uncalcined (a) and
calcined at 623 (b), and 773 K (c).
obtained by the X-ray technique. In fact, the solid was found
to be crystalline at 923 and 1073 K, but amorphous at
623 K, where it exhibited no diffraction. The uncalcined
solid was crystalline in nature [consistent with the formula yielded crystalline magnesium orthophosphate (farrington-
Mg (PO ) ) 8H O] and exhibited very weak XRD signals ite phase) in every case (d"!0.5 ppm).
3 42 2
at 773 K corresponding to farringtonite Mg (PO ) on an
3 42
amorphous background. 1
H MAS NMR Spectroscopy
Figure 6 shows the 31P MAS NMR spectra for the
magnesium oxide—magnesium orthophosphate system Owing to the dehydroxylation undergone by the solids
studied. Obviously, the intensity of the 31P signal decreased studied on raising the calcination temperature, the 1H MAS
with decreasing phosphorus content in the sample; however, NMR study was carried out only on the solids that were
the uncalcined sample always exhibited a signal at 1.1 ppm, pretreated in a stove at 393 K and those calcined at 623 and
which, consistent with the XRD results, can be ascribed to 773 K. The signals of the probe and rotor are negligible.
the phosphorus in Mg (PO ) ) 22H O. The signal for the The spectrum for the uncalcined MgO solid (Fig. 7)
3 42 2
solids calcined at 623 K appeared at 0.5 ppm, which suggests exhibited two signals, centered at 5.2 and 1.2 ppm. The
the transformation into an amorphous phase identical with former, which was the stronger, based on its downfield shift,
that formed in solid MgP. However, in these systems, MgO can be assigned to OH groups in Mg(OH) , where brucite
2
increases the crystallization temperature because the three layers are still hydroxylated. On the basis of the results of
samples remained completely amorphous at 773 K—unlike Brunet and Schaller (23), the signal centered at 1.2 ppm can
pure MgP, which exhibited some crystallinity at that tem-
perature. The crystalline phase was distinguishable in the
samples calcined at 823 K; however, there was still some
amorphous phase, the content of which increased with in-
creasing magnesium content in the system. Table 2 gives the
proportions of the two phases, calculated by Gaussian de-
convolution of the spectra. Finally, calcination at 1073 K

(b)

TABLE 2
Proportions of Amorphous and Crystalline Orthophosphate
in Mg100 – xPx Systems Calcined at 923 K
(a)
Sample % Amorphous % Crystalline

MgP 0.0 100.0

Mg P 49.4 50.6
75 25
Mg P 27.1 72.9
50 50 FIG. 8. 1H MAS NMR spectra for solid MgP calcined at 623 (a), and
Mg P 10.8 89.2
25 75 773 K (b).
 MgO—Mg (PO ) SYSTEMS
3 42
101

FIG. 9. 1H MAS spectra for solid Mg P calcined at 623 (a), 773 (b), and 923 K (c).
100~x x

be assigned to hydroxyl groups bonded to the second layer signals are thus consistent with those of amorphous magne-
of magnesium atoms. Calcining solid MgO at 623 K con- sium described earlier; however, those for MgO were not
siderably decreased the signal at 5.2 ppm as a result of the observed owing to its low content in the sample. The solid
partial dehydroxylation observed at that temperature, by remained amorphous on calcination at 773 K, consistent
which part of Mg(OH) became MgO. Raising the calcina- with the XRD and 31P MAS NMR results; also, the spec-
2
tion temperature to 773 K led to complete dehydroxylation trum was essentially identical with the previous one. The
of Mg(OH) , which was converted into periclase MgO; the spectra for solid Mg P were similar to the previous ones
2 50 50
spectrum contained three signals, at 3.6, 0.5, and !1.1 ppm, except that the signal at 5.2 ppm, corresponding to
the last of which was the strongest and must correspond to Mg(OH) , was apparent in the sample calcined at 623 K.
2
‘‘basic’’ OH groups, judging by its upfield shift. The signal at The spectra for solid Mg P exhibited well-defined signals
75 25
0.5 ppm can be assigned to ‘‘less basic’’ OH groups. Finally, for Mg(OH) ; that at 5.2 ppm was the strongest in the
2
the downfield shift of the signal at 3.6 ppm suggests the spectrum for the uncalcined sample.
presence of hydrogen bonds involving these surface hy-
droxyl groups. CONCLUSIONS
Figure 8 shows the spectra for the MgP samples calcined
at 623 and 773 K. The former contained two signals, The results obtained in this work show that the synthesis
centered at 1.2 and 3.1 ppm, that reveal the presence of two of magnesium oxide—magnesium orthophosphate systems
types of OH. On the basis of available data for AlPO , the produces solids with separate MgO and Mg (PO ) phases.
4 3 42
signal at 3.1 ppm can be assigned to tetrahedral P—OH Calcining the solids has major effects on their structure, as
groups. On the other hand, the signal at 1.2 ppm can be does the MgO content (which increases the crystallization
ascribed to Mg—OH groups. The solid calcined at 773 K temperature of magnesium orthophosphaste). The 31P
gave three signals in the 1H MAS spectrum, viz., the pre- NMR results confirmed the XRD findings. The former tech-
vious two and a third one centered at !0.1 ppm. This nique allows the proportions of crystalline and amorphous
confirms that the structure of the material still contained phases present in solids calcined at different temperatures to
some amorphous phase. On the basis of its upfield shift, the be quantified. Also, 1H MAS NMR provides qualitative
signal at !0.1 ppm can be ascribed to ‘‘basic’’ Mg—OH information on the nature of surface proton sites active in
groups. acid—base catalytic processes.
Finally, Fig. 9 shows the 1H MAS NMR spectra for
the Mg P systems. Calcination of solid Mg P at ACKNOWLEDGMENTS
100~x x 25 75
623 K resulted in the appearance of signals at 1.2 and
3.1 ppm—the latter was a shoulder of the former and was The authors express their gratitude to the Spanish DGICyT for financial
better distinguished from the shape of the first ssb. These support awarded for the realization of this work as part of Project
102 ARAMENDÏA ET AL.

PB92-0816 and to the staff of the Nuclear Magnetic Resonance Service of 11. M. A. Aramendı́a, V. Borau, C. Jiménez, J. M. Marinas, and F. J.
the University of Córdoba for their invaluable assistance in recording the Romero, J. Colloid Interface Sci. 179, 290 (1996).
NMR spectra. 12. M. A. Aramendı́a, V. Borau, C. Jiménez, J. M. Marinas, F. J. Romero,
J. A. Navı́o, and J. Barrios, J. Catal. 157, 97 (1995).
13. C. Dorémieux-Morin, C. Martin, J. M. Brégeault, and J. Fraisard,
REFERENCES Appl. Catal. 77, 149 (1991).
14. V. M. Mastikhin, U. V. Terskikh, O. B. Lapina, S. V. Filimonova,
1. J. M. Campelo, A. Garcı́a, D. Luna, and J. M. Marinas, J. Catal. 110, M. Seidl, and H. Knözinger, Solid State NMR 4, 369 (1995).
106 (1988). 15. L. Jhansi Laksmi, S. T. Srinivas, P. Kanta Rao, A. V. Nosov, O. B.
2. J. B. Moffat, in ‘‘Topics in Phosphorus Chemistry’’ (M. Grayson and Lapina, and V. M. Mastikhin, Solid State NMR 4, 59 (1995).
E. J. Griffith, Eds.), Vol. 10, p. 285. Wiley, New York, 1980. 16. M. Feike, R. Graf, I. Schnell, C. Jäger, and H. W. Spiess, J. Am. Chem.
3. J. M. Campelo, A. Garcı́a, D. Luna, and J. M. Marinas, J. Catal. 102, Soc. 118, 9631 (1996).
299 (1986). 17. P. Hartmann, C. Jana, J. Vogel, and C. Jägel, Chem. Phys. ¸ett. 258,
4. J. B. Peri, Discuss. Faraday Soc. 52, 55 (1971). 107 (1996).
5. T. Sakamoto, M. Egashira, and T. Seiyama, J. Catal. 16, 407 (1970). 18. C. S. Blackwell and R. L. Patton, J. Phys. Chem. 88, 6135 (1984).
6. T. Seiyama, M. Egashira, T. Sakamoto, and I. Aso, J. Catal. 24, 76 19. J. Tropp, N. C. Blumenthall, and J. S. Waugh, J. Am. Chem. Soc. 105, 22
(1972). (1983).
7. M. Ruwet, P. Ceckiewicz, and B. Delmon, Ind. Eng. Chem. Res. 26, 10 20. T. Kanazawa, T. Umegaki, and H. Kawazoe, ‘‘Proceedings of the
(1987). 1st International Congress on Phosphorus Compounds,’’ p. 107,
8. J. B. Moffat and J. F. Neeleman, J. Catal. 31, 274 (1973). IMPHOS, Rabat, 1977.
9. M. Giorgini, M. Bartolozzi, F. Morelli, and A. di Manno, Ann. Chim. 21. A. Pines, M. G. Gibby, and J. S. Waugh, J. Chem. Phys. 59, 569
(Rome) 65, 29 (1975). (1973).
10. M. Coudurier, M. V. Mathieu, M. Prettre, B. Imelik, and M. E. 22. S. R. Hartmann, and E. L. Hahn, Phys. Rev. 128, 2042 (1962).
Degeorges, Bull. Soc. Chim. Fr. 1821 (1968). 23. F. Brunet, and T. Schaller, Am. Miner. 81, 385 (1996).