This publication was processed (scanned, OCR'ed and manually corrected) by Andrew Wong. It has not yet been proofread. The original is in Solid State Ionics, 63-1993, 391-

Yingjeng James Li and M. Stanley Whittingham

Department of Chemistry and Materials Research Center, State University of New York at Binghamton, Binghamton, NY 13902-6000, USA

       We have investigated the formation and reactivity of new metastable phases using mild hydrothermal synthesis. Following the characterization of two new open structure sodium tungsten oxides, we found that by using a small cationic template we were able to synthesize a new layered structure titanium phosphate. Preliminary chemical analysis of this compound indicates that its hydrogen form has the formula TiO(OH)(H₂PO₄)·2H₂0. ³¹P MAS-NMR indicates that each phosphorus has two OH groups. This solid acid is readily swelled on reaction with organic amines.

1. Introduction

      There has been much interest in recent years in generating a rational approach to the synthesis of inorganic compounds, so that new structures with the desired optimum chemical and physical properties can be formed on demand. Intercalation chemistry, and "chimie douce" in general, is one such approach to rational synthesis. We have recently been exploring the use of mild hydrothermal synthesis, following the techniques of the zeolite chemist, to synthesize new metastable structures of well known compounds such as the sodium tungstates. By careful control of the synthesis conditions, such as the pH, it is possible to control the solid structure finally formed. Thus, when acidified sodium tungstate solutions are hydrothermally treated, a hexagonal form is found for 1.5≤pH≤2.0, whereas a pyrochlore structure is found in the pH range of 3.5 to 4.5 [1-3]. These relatively open crystalline lattices are metastable, reverting to the well-known high temperature phases on thermal treatment. We decided to extend these hydrothermal studies to use larger templating cations, such as the tetramethyl ammonium ion, to determine if this might lead to new interesting layer structures as has been found in the molybdenum phosphates [4]. Our initial effort has been focused on the tungsten oxide system [5] and on titanium phosphates; in both cases, new crystalline forms were found, and we report here some initial data on a titanium phosphate phase.

      Layered compounds that undergo intercalative redox reactions and ion-exchange reactions can be classified into three broad categories. The first category contains the single sheet structures of graphite and boron nitride. The second category comprises such compounds as the transition metal oxides and dichalcogenides, MS₂, which have readily reducible cations and have found extensive use as electrodes for energy storage, in electrochromic displays and as ionically reversible electrodes for electrochemical measurements. The third group is much less prone to undergo redox behaviour, but have an extensive ion-exchange and solvation chemistry. This category contains such materials as the beta-aluminas, many naturally occurring minerals such as mica, vermiculite and montmorillonite and synthetic transition metal phosphates, arsenates or silicates such as uranyl phosphate (HUP) [6] and zirconium arsenate (α-ZrAs) [7]. This third category of layered compounds has been extensively investigated as potential solid electrolytes and for ion exchange. Many of these materials tend to have relatively open crystalline lattices or closer packed lattices that are easily expanded.

      Several layered phosphates of the group IVB transition metals have been reported [8] and they crystallize in two general types of layered structures. Examples are α-zirconium phosphate (α-ZrP) of composition Zr(HP0₄)₂·H₂0 and γzirconium phosphate (γ-ZrP) of composition Zr(P0₄)•(H₂PO₄)•2H₂0. The a phase compound has been well studied because it was synthesized a long time ago and its structure is known [9]. By replacing the -OH group of o-phosphoric acid with an organic species -R or -OR, many derivatives of α-ZrP have been prepared [10-12]. Many members of the α-ZrP family, including α-ZrP itself, have received much attention because they show ion exchange, molecular intercalation, hydrogen conductivity, catalytic or other properties [13-15]. The other phase, γ-ZrP, has received less attention because its complete structure remains unknown, although a model has been proposed [16,17]. The titanium analog, α-TiP [18], has the same type of structure as α-ZrP. The gamma phase of titanium phosphate, γ-TiP [19], was reported to have the same type of structure as that of γ-ZrP. However, much less is known about the titanium compounds than those of zirconium. The titanium compounds have a potentially richer chemistry, because titanium exhibits an extensive redox behavior, whereas for zirconium only the +4 state is readily attained with reduction going directly in most cases to the metal. Here we report the preparation and preliminary characterization of a new layered titanium phosphate phase, TiO(OH)(H₂PO₄)•2H₂0, or named HTP (the hydrogen form of layered titanium phosphate) as it was obtained through hydrogen cation exchange of a precursor compound, named NMe₄TP.

2. Experimental

      NMe(sub)4TP was synthesized in a mild hydrothermal process. Since the parameters in a hydrothermal reaction are many, there is no unique condition for the preparation of NMe₄TP. A typical procedure is described as follows: 7.292 g (0.020 mole) of 25% tetramethylammonium hydroxide solution, 0.799 g (0.0010 mole) of titanium dioxide in the anatase form and 5.765 g (0.050 mole) of 85% orthophosphoric acid were mixed and introduced into a Teflon lined stainless steel sample preparation bomb (Parr 4744). The bomb was sealed and heated in an oven at 160°C for three days. After being cooled, the bomb was opened and its content filtered. The white polycrystalline product was washed several times with distilled water and dried in air. HTP was obtained by placing NMe₄TP in concentrated hydrochloric acid at room temperature for five days.

      A Scintag X-ray diffractometer with Cu Kα radiation was used to characterize the compounds formed. Thermogravimetric analysis was performed at 1°C/min in a nitrogen atmosphere on a Perkin-Elmer TG7 instrument. Quantitative analysis for Ti and P was carried out using a Fison's 55-7 DCP-AES.

3. Results and discussion

      The powder X-ray diffraction (XRD) pattern of the NMe₄TP material, fig. 1a, shows (00l) reflections with peaks of alternating intensity. The interlayer distance was calculated to be 11.4 Ǻ using linear regression from all the (00l) reflections. The HTP phase, obtained by immersing NMe₄TP in concentrated hydrochloric acid, had a smaller repeat distance of 10.0 Ǻ. Its powder XRD pattern, fig. 1b, also differs from that of the NMe₄TP in that the (00l) reflections have intensities decreasing monotonically with l.

      Thermogravimetric analysis (TGA) of the HTP phase shows various amount of weight loss below 120°C. The amount of weight loss depends on the condition in which the compound was stored. Thus, fully hydrated samples (materials stored under 100% relative humidity for longer than a month) show two steps of weight loss, as indicated in fig. 2. Each corresponds to one water molecule per formula unit; one molecule of water comes out below 400C and the other between 60°C and 120°C. The lattice repeat distance decreased to 8.9 Ǻ as these two water molecules were lost.

      The DCP-AES analysis indicated that HTP contains 27.1% wt. of titanium and 17.8% wt. of phosphorus in the dehydrated sample indicating that the mole ratio of Ti:P is 1:1. The ³¹P magic angle spinning solid state NMR spectrum, fig. 3, of the HTP sample has a resonance of about -7.5 ppm^#1 clearly

Fig. 1. Powder XRD pattern of (a) NMe₄TP; (b) HTP and (c) octylamine intercalated HTP.

showing that two of the oxygens in the P0₄ groups are present as OH groups free from structural framework [20,21]. This is in contrast to the α- and γ-phosphates, which show shifts at -18 and -32.5 ppm respectively as listed in table 1. Based on the above and the following observations the chemical formula of HTP is best proposed as TiO(OH)(H₂P0₄)•2H₂0: (1) ortho-phosphoric acid does not tend to condense into pyro-phosphate under hydrothermal condition at temperatures below 160°C; (2) the TGA weight loss on TGA does

Fig. 3. ³¹P magic angle spinning solid state NMR spectrum of HTP.

not correspond to thermal decomposition of N(CH3)₄⁺ suggesting that the tetramethylammonium ion has been entirely exchanged by protons during the preparation of HTP; and (3) the energy dispersive spectrum on an electron microprobe as well as precipitation with silver cation showed no trace of chloride indicating that chloride ions are not incorporated into HTP during the ion exchange in hydrochloric acid.

Similar to that of many other solid acids with layer structures, HTP was found to react readily with organic bases giving molecular intercalates. Thus, for example n-alkyl amines with carbons numbers be-

Table 1

³¹P NMR shifts in titanium phosphates.

	31P NMR shift in ppm			Composition
	H2PO4	HPO4	P04
α-TiP γ-TiP HTP	-10.5 -7.5	-18	-32.5	Ti(HPO4)2•H20 Ti(PO4)(H2PO4)•2H2O TiO(OH)(H2P04)•2H20

tween 5 and 14, i.e.. 5≤n≤14, can be easily intercalated into HTP by direct contact of the pure amines with the solid at 75°C for around three days. The powder XRD patterns of these intercalates show (00l) reflections, as exemplified by the octylamine complex in fig. 1c. The interlayer distances of the intercalates are listed in table 2, and as shown in fig. 4 the interlayer distances increase linearly with the

Table 2

Repeat distances of n-alkylamine intercalated HTP.

n-alkylamine	Interlayer distance (Ǻ)
amylamine hexylamine heptylamine octylamine nonylamine decylamine dodecylamine tridecylamine tetradecylamine	20.29 22.21 24.76 26.40 28.96 30.58 34.54 36.90 38.92

Fig. 4. Relation between n-alkylamines and the interlayer distances of n-alkylamine intercalated HTPs.

number of the carbons in the amine. This degree of expansion is consistent with a bilayer amine conformation with the chains tilted at an angle of 54.9° with respect to the layer basal plane.

The titanium to phosphorus ratio as well as the MAS-NMR and XRD results clearly indicate that HTP is a new layered titanium phosphate phase that differs from both α-TiP and γ-TiP. As HTP possesses two active sites on each phosphorus atom, it promises the possibility of preparing pillared derivatives by substituting -R or -OR for one or more hydrogen(s) on H₃P0₄ during the synthesis of HTP. We recently prepared a compound with an interlayer distance of about 15 Ǻ using phenylphosphinic acid instead of ortho-phosphoric acid. Therefore, a new family of layered titanium phosphate can be expected and applications on them might follow. We are presently investigating more of the fundamental properties of HTP and proposing a model for its structure.

^#1 In fact, two peaks, at -6.47 and -8.39 ppm, were observed. This might be explained as due to chemically similar but crystallographically inequivalent phosphorus atoms.

Acknowledgement

We appreciate the fruitful discussions with Professors G. Alberti of Perugia University and M.E. Thompson of Princeton University. Thanks are also due to the help from M.E. Thompson for the ³¹P MAS, solid state NMR measurement, and D.R. Naslund for the use of the DCP-AES. This work was supported in part by the National Science Foundation, DMR-8913849, and the Petroleum Research Fund, administered by the American Chemical Society.

References

[1]   K.P. Reis, A. Ramanan and M.S. Whittingham, Chem. Mater. 2 (1990) 219.

[2]   K.P. Reis, A. Ramanan and M.S. Whittingham, J. Solid State Chem. 96 (1992) 31.

[3]   J. Guo, K.P. Reis and M.S. Whittingham, Solid State Ionics 53-56 (1992) 305.

[4]   R.C. Haushalter, K.G. Strohmaier and F.W. Lai, Science 246 (1989) 1289.

[5]   J. Guo, work in progress.

[6]   V. Pekarek and M.J. Benesova, Inorg. Nucl. Chem. 26 (1964) 1743.

[7]   E. Torracca, U. Costantino and M.A.J. Massucci, Chromatogr. 30 (1967) 584.

[8]   A. Clearfield, Comm. Inorg. Chem. 10 (1990) 89.

[9]   J.M. Troup and A. Clearfield, Inorg. Chem. 16 (1977) 3311.

[10]  G. Alberti, U. Costantino, S. Allulli and N.J. Tomassini, Inorg. Nucl. Chem. 40 (1978) 1113.

[11]  M.B. Dines and P.M. DiGiacomo, Inorg. Chem. 20 (1981) 92.

[12]  D.A. Burwell and M.E. Thompson, Chem. Mater. 3 (1991) 730.

[13]  A. Clearfield, ed. in: Inorganic Ion Exchange Materials (CRC Press, Boca Raton, Fl, 1982).

[14]  G. Alberti and U. Costantino, in: Intercalation Chemistry, eds. M.S. Whittingham and A.J. Jacobson (Academic Press, New York, 1982).

[15]  M.B. Dines, P.M. Digiacomo, K.P. Callahan, P.C. Griffith, R.H. Lane and R.E. Cooksey, in: Chemically Modified Surfaces in Catalysis and Electrocatalysis, ed. J.S. Miller, ACS Symp. Ser. 192 (Washington, D.C., 1982).

[16]  S. Yamanaka and M.J. Tanaka, Inorg. Nucl. Chem. 41 (1979) 45.

[17]  A.N. Christensen, E.K. Andersen, I.G. Krogh Andersen, G. Alberti, M. Nielsen and M.S. Lehmann, Acta Chem. Scand. 44 (1990) 865.

[18]  M.A. Weiss and E. Michael, Z. Naturforsch. B 22 (1967) 1100.

[19]  S. Allulli, C. Ferragina, A. La Ginestra, M.A. Massucci and N.J. Tomassini, Inorg. Nucl. Chem. 39 (1977) 1043.

[20]  M.J. Hudson, A.D. Workman and R.J.W. Adams, Solid State Ionics 46 (1991) 159.

[21]  N.J. Clayden, Chem. Soc. Dalton Trans. 8 (1987) 1877.

This page was last updated on 14-Feb-2001 by Arne Hessenbruch.