J. MATER. CHEM., 1991, 1(3), 375-379 High-resolution Solid-state and "'Sn Magic-angle Spinning Nuclear Magnetic Resonance Studies of Amorphous and Microcrystalline Layered Metal(iv) Hydrogenphosphates Michael J. Hudson* and Andrew D. Workman Department of Chemistry, University of Reading, Whiteknights, P.0. Box 224, Reading, Berkshire RG62AD, UK High-resolution solid-state 31Pand "'Sn magic-angle spinning nuclear magnetic resonance (MAS NMR) and Mossbauer spectroscopies have been used to probe the structure of amorphous and microcrystalline layered hydrogenphosphates and some of their intercalation compounds. A linear relationship between 6('lP) and the Allred-Rochow electronegativity of the metal atom was observed. In the case of both a and y forms of titanium and zirconium hydrogenphosphates, a linear relationship was seen between the isotropic chemical shift (aiso) and the number of bridging P-0-M oxygens (connectivity) of the phosphate groups.In a further study of the intercalation compounds of tin(1v) hydrogenphosphate monohydrate (SnP), the NMR data indicated there to be more electron transfer in the propylamine intercalation compound than is the case with €-N,N'-diethylbut-2- ene-l&diamine (NNBD) and that there is more electron transfer in the case of NNBD than in the ammonium intercalation compound. A linear relationship was observed when 6(11'Sn) was plotted as a function of 6(31P). The Sn Mossbauer spectrum of the host material was indicative of ionic character around the metal atom and similar to that of SnF,.Keywords: Magic-angle spinning nuclear magnetic resonance spectroscopy; Mossbauer spectroscopy; Intercal- ation; Metal( !v) hydrogenphosphate There is currently much interest in layered hydrogenphos- phates and their intercalation compounds. Unfortunately, detailed structural studies are difficult to carry out because the materials are often microcrystalline or amorphous. X-Ray diffraction gives information principally concerning the inter- layer spacing (dOo2).Clearly other probes for structural analy- sis are urgently needed. High-resolution solid-state 31Pand I19Sn MAS NMR and Mossbauer spectroscopies offer such additional technique^.'^^ Previous studies4 on a-tin(Iv) hydrogenphosphate mono- hydrate [a-Sn(HP04)2.H20] have indicated that it is a layered compound which has an interlayer spacing of 0.78 nm and is structurally similar to a-zirconium hydrogenphosphate, (ZrP).'g6 The presence of an undissociated P-0-H group has been established from incoherent inelastic neutron scattering data.7 Similarly, studies on y-Zr(H2P04)(P04) 2H20 (y-ZrP)8*9 indicate that this is indeed the molecular formula and that the compound possesses a layered structure with two different types of phosphorus.Studies on the iso- morphous compound, y-Ti(H2P04)(P04)-2H20 (y-Tip) con- firm that this has the same layered structure as y-ZrP." Metal hydrogenphosphates have been of interest in the past for their proton conducting behaviour," ion-exchange behav- iour, catalytic properties and their use in sensors.Various mono- and poly-amines have been intercalated and studied for their ion-exchange properties and use as proton conduc- tors. The amine intercalation compounds in question here have been studied in this respect and are well ~haracterised,~ but they have not been studied previously with both 31Pand '19Sn MAS NMR. The structures of the NNBD and propyl- amine intercalates (SnP-NNBD and SnP-PrA) are represented by Fig. 1 and 2: though recent work concludes that the alkyl chains of propylamine abut one another and do not overlap.12 Solid-state 31PMAS NMR provides a probe into the elec- tronic environment of the phosphorus and hence the relative degree of deprotonation of the phosphate group by the guest amine, a matter which has previously been an item of di~pute.~ A new intercalation compound of Ru(NH~)~C~~ with SnP is Fig.1 Idealised structure of the SnP-NNBD intercalation compound presented here, the result of work on removing the radio- nuclide lo6Ru from power-station effluent stream^.'^ The new material contains a single phase and is microcrystalline. The dis0 observed by MAS NMR is related to the electronic configuration of the metal centre, and so the presence of "'Sn provides a probe for observing the changing interactions within the layers upon intercalation of the guest species. Multinuclear solid-state MAS NMR can also be used for the comparison of structurally related materials.' Tin Mossbauer provides another method of obtaining complementary infor- mation concerning layered structures related to SnP.Experimental SnP was prepared by the method of Costantino and Gas- peroni,l4 ZrP by the method of Clearfield and Stynes" and Fig. 2 Idealised structure of the SnP-PrA intercalation compound TiP by that of Alberti et ~2.'~y-ZrP was prepared by the method of Clearfield et a/.,'' and y-TiP was prepared by the method of Alberti et ul." The amorphous tin, zirconium, and titanium hydrogenphosphates were prepared using the first step of the respective methods for preparing the crystalline compounds, that is by precipitating the metal hydrogenphos- phate by addition of the metal(1v) chloride to orthophosphoric acid and then drying in uucuoat 60 "C.The amine intercalation compounds were made by the method of Hudson and Rodrig- uez-Castellon.l8 The ammonium intercalation compound was prepared by contact of the host material with concentrated aqueous ammonia, containing three times the maximum exchange capacity of the ion exchanger. The product was then dried at 60 "C in uacuo and stored in a desiccator. Analysis using powder X-ray diffraction (using a Spectrolab series 3000 X-ray diffractometer) and MAS NMR showed no evidence of hydrolysis. CHN analysis gave the formula to be Sn(HP04)o .2( PO,) 1.8(NH2)1.8. Hexaammineruthenium(I1) dichloride was prepared by the method of Fergusson and Love.l9 Its intercalation compound was prepared by shaking SnP for 20 or 180 min with a degassed aqueous solution containing 300% of the total ion-exchange capacity of SnP.The final product was dried under vacuum at 110 "C. This compound was shown by powder X-ray diffraction using the doo2 reflection and by thermogravimetric analysis to be a single phase. The initial and final ruthenium concentrations in solution were determined by atomic absorption spectropho- tometry (using a Perkin-Elmer 1 lOOB atomic absorption spectrophotometer) with a diluent containing 10% HC1 and 0.5% lanthanum trichloride as masking agents, and showed that 100% of the exchange sites had been used. The 31PMAS NMR spectra were recorded at 121.4 MHz using a cross-polar pulse sequence (spectral bandwidth 100 kHz, relaxation delay 60 s, contact time 1.0 ms, spin rate cu.5000 Hz) and gated decoupling. The 'I9Sn MAS NMR spectra were recorded at 11 1.87 MHz, a cross-polar pulse sequence (spectral bandwidth 200 kHz, relaxation delay 0.5 s, contact time 5.0 ms, spin rate cu. 5000 Hz) with flip-back. These were run by the SERC NMR service of the Industrial Research Laboratories at Durham. Peak widths were meas- ured throughout as FWHM. The "'Sn NMR spectra were referenced to Sn(CH3)4, while the 31P NMR spectra were referenced to orthophosphoric acid. Sn Mossbauer analysis was carried out at the Demokritos National Research Centre for Physical Sciences, Athens, Greece. J. MATER. CHEM., 1991, VOL. 1 Results and Discussion Connectivity of Phosphate Groups The 31PNMR were clear and well resolved in the crystalline materials.Fig. 3 shows the linear relationship between the connectivity of the phosphorus (Q") and the isotropic chemical shifts for a-and y-zirconium (ZrP) and titanium hydrogen- phosphates (Tip). The nomenclature is the same as that used to describe condensed silicates and is characterised by the number of bridging oxygens of the phosphate group, n. In the case of a-ZrP, each phosphorus is connected to three metal centres via bridging oxygens and so has a connectivity of three. In the case of y-ZrP, there are equal amounts of two different phosphate groups, one connected to four Zr ions, Q4, while the other is a dihydrogenphosphate connected to two Zr centres and so is Q2." The direction of the gradient in Fig.3 is indicative of increasing summed P-0 bond strength2 with increasing connectivity. This technique was used to confirm that y-TiP has the same type of layered structure as y-ZrP. Electronegativity of the Metal Ion The isotropic chemical shifts of amorphous and a-structured metal(rv) hydrogenphosphates (M =Sn, Ti, Zr) plotted as a function of Allred-Rochow electronegativity of the metal centre are shown in Fig.4. It can be seen that there is a correlation between the electronegativity of the metal centre and isotropic chemical shift. The direction of the gradient shows increasing deshielding of the phosphorus attributable to growing .n electron density and the summed phosphate bond strength'i2 with decreasing electronegativity. The large differences in the chemical shifts of the phosphate groups with different connectivities have already been established in pre- vious =I \\ ~ -30 I 0 -35 1 2 3 4 5-Q Fig.3 6i,0(31P)expressed as a function of the connectivity of the phosphate groups for zirconium (+) and titanium (0)hydrogenphos-phates -251 4 I I I I 1 1 ,1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 electronegativity Fig.4 Si,0(3'P) expressed as a function of the electronegativity of the metal ion for a-(+) and amorphous (0)hydrogenphosphates J. MATER. CHEM., 1991, VOL. 1 Amorphous Compounds Using the above method of assignments for the connectivities of the phosphate groups, the amorphous compounds consist of predominantly Q3 phosphate groups, which are associated with the a-type structure.There are differences between the spectra, however, which can be divided into two categories, the width of the signal and the number of signals. The tin hydrogenphosphate spectrum (Table 1) consists of a single peak, ca. 26 ppm wide. The breadth is indicative of a low degree of crystallinity rather than hydrolysis. There was no evidence of a y-form of SnP. Only one resonance was observed in the "'Sn MAS NMR spectrum, indicating that there was interestingly no hydrolysis to Sn02 and this was confirmed using Mossbauer spectroscopy as discussed later. The spectrum of amorphous zirconium hydrogenphosphate (Fig. 5) consists of two peaks and a shoulder. The first (smaller) peak lies at -12.71 ppm, the second (and principal peak) at -20.35 ppm and the shoulder at -26 ppm.These values suggest that the material consists principally of the a-form with some y-phase as the minor component. The estimated breadth of the central peak is 17ppm. Again, the spectrum of amorphous titanium hydrogenphosphate (Fig. 6) consists of three resonances, -16.02, -18.76 and -30.22 ppm, though the bandwidth is now much reduced to ca. 5 ppm, indicating a higher degree of crystallinity. The peak at the highest resonance, -16.02 ppm is unusual, as it does not appear low enough for a Q2 phosphate. It may be that this is a polyhyd- rated hydrated Q3 phosphate with hydration leading to a reduction in the chemical shift. Conversely, as discussed later with the NNBD intercalation compound, the chemical shift in hydrophobic regions is greater than that in hydrated regions.The X-ray powder diffraction pattern of this com- Table 131P MAS NMR of some tin(1v) hydrogenphosphate com-pounds compound isotropic chemical shift (ppm) amorphous SnP -13.74 (26") SnP-Pra -7.35 SnP-NH: -6.94 -11.17 -14.19 SnP-NNBD -10.28 -15.35 Denotes width in ppm taken as FWHM. -Tv~l-n-r-~-~--10 -20 -30 -40 6 (PPm) Fig.6 31P MAS NMR spectrum of amorphous titanium hydrogen- phosphate pound shows no structure at This can be rationalised on the basis of the compound having a high degree of order but only over a very short range, such that is undetectable by X-ray diffraction. This, therefore, is a truly microcrystalline compound. In the two spectra of multiple resonances, there- fore, most peaks can be assigned to a particular phosphate connectivity, the closest to zero being Q2, the next Q3 and the last Q'.31PMAS NMR of some Intercalation Compounds of SnP The 31PNMR spectrum of crystalline SnP is shown in Fig. 7 to be clear and well resolved, as is the case with the other spectra, which are listed in Table 1. The principal peak occurs at -13.6 ppm and is assigned to phosphorus with an undis- sociated P-0- H group.7 No other phosphate connectivities are seen within this spectrum and comparison with the spectrum of the amorphous compound shows a decrease in linewidth which is related to the increase in crystallinity of the material. The intercalation compound with propylamine analyses as Sn(CH3CH2CH2NH2)2(HP04)2and has a similar structure to SnP, but the dOo2interlayer spacing has been increased to 1.672 nm. The 31PMAS NMR spectrum shows only one type of phosphorus with a chemical shift of -7.34 ppm.The single peak confirms that all of the Q3phosphate groups are involved with bonding to the guest molecule as implied by the 2: 1 amine to tin ratio. It has been shown1V2 that in 31P MAS NMR the isotropic chemical shift moves upfield as the P-0 I""I""l""~""I"I 1 I I I 0 -10 -20 -30 100 0 -100 6 (PPm) 6 (PPW Fig. 531P MAS NMR spectrum of amorphous zirconium hydrogen- Fig.7 31P MAS NMR spectrum of crystalline a-tin hydrogen-phosphate phosphate bond strength increases, corresponding to a decreased para- magnetic term.There is some degree of proton transfer from the phosphate group to the amine, the decrease of 6.5ppm being associated with an increase in negative charge on P-0"-(~21).However, if the free electron density were to result in an increase in the double-bond character of the phosphate group, an opposite movement in the chemical shift would be seen. The movement observed can possibly be explained on the basis of charge localisation on the oxygen, encouraged by hydrogen bonding between the protonated amine and the layer. The phosphorus is, therefore, more shielded and the peak moves accordingly. The spectrum of the ammonium intercalation compound (SnP-Amm) shows three resonances with the major peak at -1 1.1 7 ppm.The small peak at -6.95 ppm is attributable to either a phosphate bonded to both an ammonium ion and water or a diammoniated phosphate, while that at -14.19 ppm is due to an unexchanged phosphate group. The latter has a chemical shift lower than that of the parent tin hydrogenphosphate. This is due to a different extent of hydration with the associated change in hydrogen bonding. The pK, of the ammonium ion is 9.3.21 The extent of proton transfer from the host compound to the guest is larger than is the case with water and the resonance moves for the same reasons as given above. The move, however, is not as great as is the case of SnP-PrA, which is a stronger base (pK,= 10.71 for propylamine).21 The "N MAS NMR spectrum of the ammonium intercalation compound shows only one peak, indicative of only one nitrogen environment.With respect to the intercalation compound with the diamine NNBD as guest molecule, structural studies have already been published.22 The compound analyses as Sn(NNBD)o,73(HP04)2.Like the propylamine intercalation compound, the structure is closely related to that of SnP but with an interlayer spacing of 1.35 nm. Since the ratio of NNBD to Sn is less than unity, some of the 0-H groups are not involved with binding to the amine groups. There are two separate peaks in the spectrum with chemical shifts of -10.3 and -15.4 ppm, confirming the two phosphate environ- ments. This is intermediate between the ammonium intercal- ation compound and SnP, suggesting that the protonation is greater than in SnP but less than in the ammonium or propylamine intercalation compounds, but not as expected from the pK, of the base (the first pK, of NNBD is greater than 11.8).This could be connected to the relatively bulky groups on the diamine holding the nitrogen away from the active site on the layer (area 21.4 A2).23There are parallels in solvent extraction studies and similar phenomena are observed in the active site of enzyme^.^^,^' The second peak at -15.4 ppm is interesting as it is lower than the shift in SnP itself. This low chemical shift is probably due to the shielding of the phosphate active site by the bulky alkyl groups of the diamine keeping water away from the 0-H group by the creation of a hydrophobic environment, as shown in Fig.2. Metal hydrogenphosphates were found26 to exchange radioactively labelled phosphate rapidly from aque- ous solution and the pendant P-0 bonds bend out to accommodate large guest species leaving some hydrophobic regions. l19Sn MAS NMR of SnP Intercalation Compounds Fig. 8 is a plot of 6iso(119Sn) as a function of C~~,,(~~P). It can be seen that there is a linear correlation between these two terms. As the charge on the phosphate group increases so 6(l19Sn) falls. This movement in the isotropic chemical shift is attributable to increasing shielding of the tin nucleus by J. MATER. CHEM., 1991, VOL. 1 -a "'Sn isotropic chemical shift (ppm) m Fig.8 6,,(119Sn) expressed as a function of 6i,,(31P) in reference to Sn1V(P04)2X2,[X =H, Ru(NH,),, protonated amines]: (a) SnP; (b) SnP-Ru(NH,),, 20 min intercalation time; (c) SnP-Ru(NH,),, 180 min intercalation time; (d) SnP-hexylamine intercalate; (e) SnP-Pra; (f) SnP-octylamine intercalate increasing electron density.A previous has shown that there are correlations between 6(lI9Sn) and the amount of electron density around the Sn nucleus and confirms the direction of the change. Mossbauer Spectra The Mossbauer spectrum for SnP is shown in Fig. 9; the spectra for all the intercalation compounds are identical. It can be seen that the value of 6 is -0.361 f0.005 mm s-' with a half-height width of 1.05 mm s-'. These values resemble those for SnF,28 which implies that the tin is ionic in character.By comparison, the Sn Mossbauer of Sn02 (67 0.0015 mm s-') has a larger half-height width of 2.52 mm s-. The tin Mossbauer spectrum of the hexaammineruthenium(1r) intercalation compound of SnP shows little difference from the spectrum of the host material, except a slight movement in 6 towards that of the more covalent Sn-0 bond in Sn02. The spectra for all of the other compounds were identical indicating that the principal changes may be better studied using the NMR techniques. Conclusions There are linear correlations between the chemical shift of the phosphorus and both the connectivity of the phosphate group and the electronegativity of the metal ion. The amorphous compounds are shown to consist of predominantly Q3phos-phate groups.A linear correlation was observed with 6iso('19Sn) was plotted as a function of 6i,0(31P). C .-.i! 98.47-5 97.69-96.92 .. I I I I96.141 '-' I , I , 11 J. MATER. CHEM., 1991, VOL. 1 We are grateful for the funds provided by the Department of the Environment for A. D .W. The results may be used in the formulation of Government policy, but at this stage do not necessarily represent Government policy. We also wish to thank Dr. D. C. Apperley at the SERC NMR facilities of the University of Durham. Professor G. Alberti kindly gave us the sample of y-ZrP and the a-ZrP, a-Tip, y-TiP and their amorphous counterparts were donated by Mr. R. J. W. Adams. Amorphous tin hydrogenphosphate was provided by Mr.P. Sylvester. The Sn Mossbauer spectra were run by Dr. D. Petridis of Demokritos in Athens. 10 11 12 13 14 15 16 17 G. Alberti, U. Costantino and M. L. Luciani Giovagnotti, J. Inorg. Nucl. Chem., 1979, 41, 643. M. Casciola and D. Bianchi, Solid State Ionics, 1985, 17, 287. G. Alberti, personal communication, 199 1. Directorate of Fisheries Research, Lowestoft (Annual Sub-missions), 1986. U. Costantino and A. Gasperoni, J. Chromatogr., 1970, 51, 289. A. Clearfield and J. A. Stynes, J. Inorg. Nucl. Chem., 1964, 26, 117. G. Alberti, P. Cardini Galli, U. Costantino and E. Torracca, J. Inorg. Nucl. Chem., 1967, 29, 571. A. Clearfield, R. H. Blessing and J. A. Stynes, J. Inorg. Nucl. Chem., 1968,30, 2249. 18 M. J. Hudson and E. Rodriguez-Castellon, J.Incl. Phenom., 1989, 7, 301. References 19 J. E. Fergusson and J. L. Love, Inorg. Synth., 13, 1971, 208. F. Taulelle, C. Sanchez, J. Livage, A. Lachagar and Y. Piffard, J. Phys. Chem. Solids, 1988, 49, 229. A. K. Cheetham, N. J. Clayden, C. M. Dobson and R. J. B. Jakeman, J. Chem. SOC., Chem. Commun., 1986, 195. I. L. Mudrakovskii, V. P. Shmackova, N. S. Kotsarenko and V. M. Mastikhin, J. Phys. Chem. Solids, 1986,47, 335. M. J. Hudson, E. Rodriguez, P. Sylvester, A. Jiminez-Lopez and P. Olivera-Pastor, Hydrometallurgy, 1990, 24, 77. A. Clearfield and G. D. Smith, Inorg. Chem., 1969, 8,431. A. Clearfield and J. M. Troupe, Inorg. Chem., 1977, 16, 331 1. D. J. Jones, J. Penfold, J. Tomkinson and J. Roziere, J. Mol. Structure, 1989, 197, 113. N. J. Clayden, J. Chem. SOC., Dalton Trans., 1987, 77. 20 21 22 23 24 25 26 27 28 R. J. W. Adams, personal communication, 1991. Handbook of Chemistry and Physics, ed. R. C. Weast, CRC Press, Boca Raton, p. 33431. M. J. Hudson and P. Sylvester, Solid State Ionics, 1989, 35, 73. I. Fotheringham, M.Sc. Thesis, 1986, University of Reading, UK. M. J. Hudson, Hydrometallurgy, 1982, 9, 149. L. Stryer, Biochemistry, W. H. Freeman, San Francisco, 1989. A. Clearfield, personal communication, 1991. S. J. Blunden, D. Searle and P. J. Smith, Inorg. Chim. Acta, 1986, 116, L31. V. T. Goldanskii, V. Ya. Rochew, V. V. Khraphov, B. E. Dzevitskii and V. F. Sukhoverkhov, Isv. Sib. Otd. Akad. Nauk., Ser. Khim. Nauk., 1968, 22. C. Y. Oritz-Avila and A. Clearfield, J. Chem. Soc., Dalton Trans., 1989, 1617. Paper 01050361; Received 9th November, 1990