摘要:

Synthesis and characterization of a novel layered titanium silicate JDF-L1 Hongbin Du, Min Fang, Jiesheng Chen and Wenqin Pang* Department of Chemistry, Jilin University, Changchun 130023, China A novel layered titanium silicate JDF-L1 has been synthesized in an Si02-Ti02-Na20-H20-H202 template system. It is found that the synthesis of JDF-L1 requires a relatively high alkalinity and an appropriate titanium content. The JDF-L1 samples were characterized by XRD, SEM, IR, DTA-TG and adsorption. DTA-TG studies, combined with XRD and IR spectra, show that JDF-L1 is stable up to 873 K. After activation at 523 K under vacuum, the samples exhibit significant water-adsorption capacities. It is also revealed that JDF-L1 can be intercalated with alkylammonium cations. There are many natural titanium silicates occurring in the earth's crust, but only a few such as verplanckite' and zorite' have an open framework. Recently, the successful synthesis of open-framework titanium silicates ETS-4,3*4 and ETS- 10' con- tributed significantly to our understanding of the nature and chemistry of the titanium silicate family.Moreover, the well known synthetic titanium-substituted zeolites such as Ti-p,6,7 TS-18*' as well as other titanium-substituted microporouslo." and mesoporous ~ilica,'~~'~ have attracted much attention during the last decade because of their remarkable catalytic activity and selectivity in the hydroxylation of organic com- pounds with dilute aqueous hydrogen peroxide. Almost all of the titanium silicates are composed of tetrahedral SiO, and octahedral Ti06 units.In the dehydrated titanium-substituted zeolites, however, the Ti is four-coordinate and expands its coordination sphere number by interaction with adsorbates.6,8 So far there is little information in the literature on titanium silicates with five-coordinate Ti. To our knowledge, the mineral fre~noite,'~Ba2TiSi208, contains uncommon square-pyramidal TiO, polyhedra. Recently we synthesized a novel layered titanosilicate Na4Ti2Si8O2, 4H20," which we designated JDF-L1 (Jilin-Davy-Faraday layered solid no. 1). In this structure, Ti atoms are five-coordinate and each is linked to Si04 tetrahedra forming continuous sheets. Sodium cations and water molecules are occluded between the layers, with the former compensating the framework charge.It is well known that pillared or intercalated layered mate- rials are of considerable interest in the search for advanced materials. The possibility to controllably modify their chemical composition as well as pore size provides the flexibility to tailor these materials for specialized uses, e.g. as catalysts, catalyst supports, ion exchangers16-18 and composite mate- rial~.'~The successful synthesis of the layered titanium silicate JDF-L1 with a unique structure will open up many intriguing possibilities in materials chemistry, especially in catalytic con- version of organic species and in solid-state electrolytes. The present paper focuses on the synthesis and properties of JDF-L1. The influence of the synthesis parameters such as Ti02/Si02 and Na,O/SiO, ratios on the products is also described.Experimental Synthesis JDF-L1 was synthesized hydrothermally using tetrabutyltita- nate (TBOT, 98%) as the titanium source, fumed silica (98%) as the silicon source and tetrapropylammonium bromide (TPABr, Fluka) as the template. The chemical composition of the initial gel was 1.0 SiO, :xTi0, :yNa,O :0.12TPABr :2.5H202:8.0NH3: 36H20, where x =0.1-0.5 and y =0.3-0.6. In a typical synthesis, TBOT was dispersed in aqueous ammonia (25 %), and aqueous hydro- gen peroxide (30%) was added. After stirring, a clear yellow solution was formed. To the above solution sodium hydroxide and TPABr were added, followed by fumed silica.The mixture was stirred until it became homogeneous, then transferred into a Teflon-lined stainless-steel autoclave, and heated at 453 K for about 7 days. The products were recovered by filtration, washed and dried at 353 K. Characterization Powder X-ray diffraction (XRD) data were obtained using a Rigaku D/MAX-IIIA diffractometer with Cu-Ka radiation (1=0.15418 nm). A Hitachi X-650B scanning electron micro- scope was used for SEM experiments. Thermogravimetry (TG) and differential thermal analysis (DTA) were performed on a Perkin-Elmer TGA 7 thermogravimetric analyser and a DTA- 1700 differential thermal analyser, respectively, under a flow of N2 at a heating rate of 10 K min-'. IR spectra were recorded on a Nicolet 5DX FTIR spectrometer using the KBr pellet technique.Adsorption was measured isothermally on a Cahn-2000 electronic recording balance. Prior to each measurement, the as-synthesized sample was activated by heating at 453 K while maintaining a vacuum of t0.133 Pa for ca. 3 h followed by cooling to room temperature under vacuum. Chemical analysis of the crystallized products was carried out with a Leeman ICP-AES instrument. Results and Discussion Synthesis JDF-L1 can be synthesized using several amines or ammonium salts, e.g. triethylamine, tributylamine, ethylamine, TPABr, tetrabutylammonium broxide, as structure-directing agents, indicating that the nature of the agents has little influence on the formation of JDF-L1. However, the effect of the templates on the morphology of JDF-L1 is distinct.Fig. 1 shows the images of the JDF-L1 samples synthesized with different amines. With triethylamine as the template, the sample is composed of highly intergrown or twinned aggregates [Fig. 1 (a)], whereas with TPABr, square plate-like crystals approximately 25 x 25 pm2 in size [Fig. 1(b)] were obtained. The alkalinity in the gel plays an important role in the formation of JDF-L1. Fig.2 shows the influence of the Na,O/SiO, molar ratios on the crystallization of JDF-L1. At low sodium hydroxide concentration, a compound isostructu- ral with ZSM-5 (MFI) is the main crystalline product [Fig. 2(a)]. An increase in the sodium hydroxide concentration results in the formation of the layered titanium silicate JDF- J.Mater. Chem., 1996,6( 11), 1827-1830 1827 I -1 Fig. 1 Scanning electron micrographs of JDF-L1 synthesized with (a) triethylamine and (b) TPABr as template 10 20 30 40 2Ndeg rees Fig. 2 XRD patterns of JDF-L1 showing the influence of Na,O/SiO, ratios on the crystallization. Curves (a)-(d) correspond to the Na,O/SiO, ratios of 0.24, 0.48, 0.60 and 0.72, respectively. L1. The suitable Na,O/SiO, molar ratios for JDF-L1 range from 0.35 to 0.6. A further increase in alkalinity leads to poorer crystallinity of the sample [Fig. 2(c)], and the mineral zorite is crystallized with a small amount of JDF-L1 [Fig. 2(d)] when the Na20/Si02 molar ratio reaches 0.72. As shown in Fig. 3, the titanium content in the reaction mixture also affects the crystallization of JDF-L1.A low titanium content in the gel usually results in the formation of unknown dense phases [Fig. 3(d)]. At the highest titanium content, however, JDF-L1 is accompanied by a considerable amount of amorphous material and fully crystalline samples cannot be obtained by increasing the crystallization time [Fig. 3(a)]. Highly crystalline JDF-L1 products are favourable only in the narrow Ti02/Si02 ratio range 0.1-0.35 when the ratio of Na,O/SiO, is fixed at 0.48. X-Ray powder diffraction The X-ray powder diffraction pattern of the as-synthesized JDF-L1 is presented in Fig. 4. The product is phase-pure and all of the diffraction lines in the pattern can be indexed in a tetragonal symmetry.The unit-cell parameters are a =b = 0.7374 nm, c =1.0710 nm, in agreement with the results obtained previou~ly.~~ 1828 J. Muter. Chem., 1996, 6(11), 1827-1830 I. I . I. I. I 10 20 30 40 28/degrees Fig. 3 XRD patterns of JDF-L1 showing the influence of TiO,/SiO, ratios on the crystallization. Curves (a)-(d) correspond to the TiOJSiO, ratios of 0.50, 0.25, 0.18 and 0.10, respectively. I1 10 20 30 40 50 2e/degrees Fig. 4 X-Ray diffraction pattern of the as-synthesized JDF-L1 Like most clays, such as kaolin and montmorillonite, the interlamellar alkali-metal ions in JDF-L1 are exchangeable. JDF-Ll can be intercalated with an aqueous solution of nonyl- ammonium salt. As shown in Fig. 5, the interlayer d-spacing of the pillared JDF-L1 increases significantly.Since alkyl chains grow by ca. 0.127 nm per added carbon, the expanded distances (ca. 3.4 nm) suggests that the nonylammonium ions form double layers with the NH3+ groups pointing toward the basal titanium silicate layers. The successful pillaring of JDF-L1 provides an effective route to utilize it for chemical or catalytic reactions. Preliminary studies on the organic ion-expanded lamellar JDF- L1 shows promising results. Further investigation relevant to the intercalation is under way. IR spectroscopy The FTIR spectra of the as-synthesized and thermally treated JDF-L1 samples are shown in Fig. 6. The spectrum of the as- synthesized sample is completely different from those of the titanium silicate zeo1ites20v21 and mineral ~orite,~ but rather similar to that of the fresnoite Ba2TiSi208,22 probably because of some analogies between the two structures, especially due to the existence of pyramidal TiOS groups.Gabelica-Robert and Tarte2’ investigated the IR and Raman spectra of fresnoite and fresnoite-like pyrosilicates on the basis I 0 10 20 30 40 2BIdegrees Fig. 5 X-Ray diffraction patterns of (a) as-synthesized JDF-L1 and (b) nonylammonium-pillared JDF-L1 1400 1000 800 600 400 wavenum berIcm -l Fig. 6 IR spectra of JDF-L1 (a) as-synthesized, and heated in air at (b) 523 K, (c) 823 and (d) 923 K, respectively of a factor group analysis and with the help of 28129Si and 46/50Tiisotopic shift. They attributed the 1400-900 cm-' region of the spectra to the stretching vibrations of SiO groups and 900-800cm-1 peaks to both the v,,(Si03) and the y(Ti-0) vibrations. According to the framework vibration models of zeolites23 and fresnoite,,, the IR absorption spectrum of the as-synthesized JDF-L1 can be divided into the following three main vibration ranges.The peaks occurring in the region 1400-950 cm-' are assigned to the asymmetric stretching vibrations of Si-0-Si and/or Si03 groups. The bands between 600 and 850 cm-' are associated with symmetric stretching vibrations of SO4. The ring and bending vibrations are responsible for the absorption at 400-600 cm-'. By com- parison with the absorption of fresnoite, the band at 888 cm-' in JDF-L1 should be related to the Ti05 group.Structure analysis'' reveals that in JDF-L1 there exists a short Ti-0 bond in a discrete Ti05 groups, which is probably associated with the IR absorption in the 900-800 cm-' range. The IR spectra of the JDF-L1 samples treated at different temperatures presented in Fig. 6 show that below 823 K the structure of JDF-L1 remains. However, it is changed above 923 K. This is consistent with the following results of thermal analysis. Thermal properties To investigate the thermal properties of the titanium silicate JDF-L1, DTA-TG analyses in a flowing N, atmosphere were carried out and the results are shown in Fig. 7. The DTA curve has two endothermic effects occurring at ca. 433 and 543 K, respectively. This is in good agreement with the TG results.The TG curve can be divided into two corresponding stages ranging from 373 to 453 K and 453 to 673 K. The correspond- ing mass losses are 2.8 and 5.5%, respectively, with a total loss of 8.3%. Element analysis shows that the as-synthesized JDF- L1 occludes no organic molecules. In addition, the total mass loss is consistent with the calculated value (8.6%) of four H20 molecules per formula of the compound. These suggest that the two endothermic effects are assigned mainly to the evol- ution of the occluded water. Upon removal of the occluded water from 423 to 523 K, JDF-L1 begins to lose its crystallinity, as shown in Fig. 8. However, its structure remains until 873 K in terms of X-ray diffraction. When the temperature reaches 923 K, however, JDF-L1 collapses and converts into a dense, non-zeolite phase, although neither a corresponding thermal effect nor a mass loss is evident in the TG and DTA curves.I I 5.0 100 hs Y % 90 2.5 z d 1 I I80 LL "0 303 423 543 663 783 903 TIK Fig. 7 TG-DTA curves for JDF-L1 10 20 30 40 E 2BIdegrees Fig. 8 XRD patterns of JDF-L1 (a) as-synthesized, and heated in air at (b) 423, (c) 523, (d) 823 and (e) 923 K for 3 h, respectively J. Mater. Chem., 1996,6( ll), 1827-1830 1829 lo I 8 h Y86 \c E4 2 I 1 I I I 1 0 0.1 0.2 0.3 0.4 0.5 0.6 PIP0 Fig. 9 H,O adsorption isotherm for JDF-L1 at 293 K Adsorptive properties The water adsorption isotherm for JDF-L1 at 293 M is shown in Fig.9. The curve corresponds well to a type I adsorption isotherm. The adsorption capacity of water is 8% at PIPo= 0.2, indicating that water molecules can readily penetrate into the internal space of JDF-L1. However, only about 2% n-hexane is adsorbed, indicating only extracrystalline adsorption on the solid surface. Conclusions A novel layered titanium silicate JDF-L1 was hydrothermally synthesized. Investigations show that the synthesis of JDF-L1 involves two main factors: Ti0,/Si02 molar ratios and the alkalinity in the gel. The amines have little influence on the formation of JDF-L1. The X-ray powder diffraction pattern reveals that JDF-L1 can be intercalated by protonated alkyl- ammonium cations.The short Ti-0 bond in JDF-Ll contrib- utes to the 888 cm-' absorption of the IR spectrum. JDF-Ll is stable up to 873 K and converts into a non-zeolite structure at about 923 K. After dehydration at 453 K, the samples adsorb ca. 8 mass% water but little n-hexane at room tempera- ture when PIP, =0.2. This work was supported by the National Natural Science Foundation of China and the Key Laboratory of Inorganic Hydrothermal Synthesis of Jilin University. References 1 A. R. Kampf, A. A. Khan and W. H. Baur, Acta Crystallogr., Sect. B, 1973,29,2019. 2 P. A. Sandomirskii and N. V. Belov, Kristallografya, 1979, 24, 1198. 3 D. M. Chapman and A. L. Roe, Zeolites, 1990,10,730. 4 H. Du, F. Zhou, W. Pang and Y. Yue, Microporous Muter., 1996, 7, 73.5 M. W. Anderson, 0. Terasaki, T. Ohsuna, A. Philippou, S. P. MacKay, A. Ferreira, J. Rocha and S. Lidin, Nature (London), 1994,367,347. 6 T. Blasco, M. A. Camblor, A. Corma and J. Perez-Pariente, J. Am. Chem. SOC., 1993,115,11806. 7 M. A. Camblor, A. Corma and J. Perez-Pariente, Zeolites, 1993, 13,82. 8 S. Bordiga, S. Coluccia, C. Lamberti, L. Marchese, A. Zecchina, F. Boscherini, F. Genoni, G. Leofanti, G. Petrini and G. Vlaic, J. Phys. Chem., 1994,98,4125. 9 M. Taramasso, G. Perego and B. Notari, US Pat., 4,410,501, 1983. 10 D. P. Serrano, H-X. Li and M. E. Davis, J. Chem. SOC., Chem. Commun., 1992,745. 11 J. S. Reddy, R. Kumar and P. Ratnasany, Appl. Catal., 1990,58, L1. 12 P. T. Tanev, M. Chibwe and T. J. Pinnavaia, Nature (London), 1994,368,321. 13 T. Maschmeyer, F. Rey, G. Sankar and J. M. Thomas, Nature (London),1995,378,159. 14 P. B. Moore and S. J. Louisnathan, Z. Kristallogr., 1969, 130,438. 15 M. A. Roberts, G. Sankar, J. M. Thomas, R. H. Jones, H. Du, J. Chen, W. Pang and R. Xu, Nature (London), 1996,381,401. 16 F. Cavani, F. Trifiro and A. Vaccari, Catal. Today, 1991,11,173. 17 F. Figueras, Catal. Rev. Sci. Eng., 1988,30,457. 18 T. J. Pinnavaia, Science, 1983,220,365. 19 G. A. Ozin, Adv. Muter., 1992,4,612. 20 A. Thangaraj, M. J. Eapen, S. Sivasanker and P. Ratnasamy, Zeolites, 1992, 12,943. 21 J. S. Reddy and R. Kumar, Zeolites, 1992, 12,95. 22 M. Gabelica-Robert and P. Tarte, Phys. Chem. Mineral., 1981, 7,26. 23 E. M. Flanigen, H. Khatami and H. A. Szymanski, Adv. Chem. Ser., 1971,101,201. Paper 6/02346K; Received 3rd April, 1996 1830 J. Mater. Chern., 1996, 6(11), 1827-1830

ISSN:0959-9428    

DOI:10.1039/JM9960601827

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

The structure of the calcined aluminophosphate ALPO,-5 determined by high- resolution X-ray and neutron powder diffraction AsiloC J. Mora,"yb Andrew N. Fitch,"." Michael Cole," Rajan Goyal," Richard H. Jones; HervC Jobicd and Stuart W. Carr" 'Department of Chemistry, Keele University, Stagordshire, UK ST5 5BG bDepartamento de Quimica, Facultad de Ciencias, Universidad de Los Andes, La Hechizera, Mirida 5101, Venezuela "ESRF, BP 220, 38043 Grenoble Cedex, France dInstitut de Recherches sup la Catalyse, 2, Avenue Albert-Einstein, 69626 Villeurbanne Cedex, France "Unilever Research Port Sunlight Laboratory, Quarry Road East, Bebington, Wirral, UK L63 3J W Rietveld refinements of the structure of calcined ALP04-5 have been performed using synchrotron and neutron powder diffraction data.The high-resolution synchrotron X-ray diffraction pattern shows splitting of some peaks, which indicates that the symmetry is lower than the hexagonal cell observed in earlier studies. The neutron powder pattern is of lower resolution and no splitting is apparent, but the different scattering lengths of A1 and P, and the higher scattering power at high angle, provide a sound test for the different models investigated. Good fits of the neutron and synchrotron data were obtained in space group Pcc2 with cell parameters a =13.794( 3) 8,b =23.900(6)8,c =8.4168(6) A.In this space group there is a strict alternation of A1 and P in the framework, as the 31Pand 27Al MAS NMR spectra indicate, and two crystallographically distinct 12-ring channels run parallel with the c axis.Microporous aluminophosphates have been the focus of wide- spread investigation since their first synthesis in 1982.1,2 These compounds are built up from alternating vertex-sharing AlO, and PO4 tetrahedra. Whilst some of these compounds and their hetero-substituted derivatives posses the same structure as naturally occurring zeolites, e.g. A1P04-17 and SAP04-37 are isostructural with erionite3 and faujasite,, respectively, some of them have structures for which there is no known natural counterpart. Some of these materials have frameworks with large apertures which permit the adsorption of large molecules and their use as catalysts. For example, 12-membered rings are found in DAF-1,' 14-membered rings in ALP04-8,6 and 18-membered rings in VPI-5.7 However, one of the simplest microporous aluminophosphates is ALP04-5 which was reported in the initial studies.',' Two single-crystal structure determinations have been carried out on the as-synthesised material using the hexagonal space group P~cc.~,'A powder neutron diffraction study has also been performed on a calcined sample.l0 These studies show that ALP04-5 possesses a one-dimensional 12-membered channel which is surrounded by 4- and 6-rings.Despite this success there are some problems with these analyses. In the case of the single-crystal analyses there is an A1-0-P angle close to 180°, and the temperature factor of this particular oxygen atom is rather large. One set of authors suggests that this is due to three-fold disorder.8 A serious problem is encountered in the powder neutron study on the calcined material.Here it proved impossible to develop an adequate model in the space group P~cc, and the higher symmetry space group P6/mmc was adopted. This requires that the aluminium and phosphorus atoms are distributed randomly over the tetrahedral sites, and this is clearly incom- patible with the single-crystals studies and general chemical experience. Theoretical investigations on the ALP04-5 system have suggested that symmetries lower than hexagonal might exist." Here we report how a combination of high-resolution X-ray and neutron powder diffraction has been able to unravel this problem. The ALP04-5 sample was prepared as follows.12 Pseudoboehmite (41.1 g, Catapal B from Vista Chemical Co) was added to a solution of phosphoric acid (57.5 g, 85%) and water (149 g).This was stirred with a paddle for 2 h at room temperature. To this mixture was added triethylamine (38.5 g) and the stirring was continued for 0.5 h. The mixture was transferred to a Teflon-lined autoclave and heated without stirring at 200°C for 2 h. The white microcrystalline powder was collected by filtration, washed with water (2 x 100 cm3) and dried in air at 80°C. Scanning electron microscopy (SEM) of the as-prepared material gave the typical hexagonal cylindrical-shaped crystals. The 31P and 27AlMAS NMR spectra of the as-prepared material gave single resonances at 6 -29.6 and 37.6, respectively.These agree with the published data,l2.I3 and are consistent with tetrahedral A1 (4P) and P(4A1) environments. Thus there is strict alternation of the A1 and P atoms in the framework. A refinement of the structure of the as-prepared material from high-resolution powder synchrotron diffraction data gave a satisfactory fit (R, = 11.3%, RI=6.2%, Rexp=7.0%) in the hexagonal space group P~cc, in agreement with the previous single-crystal determi- nations. The calcined ALPO,-5 was prepared by heating a thin bed of the as-prepared material at 800°C in air, which was then equilibrated at room temperature after calcination. The 31Pand 27Al MAS NMR spectra of the calcined sample give single broader resonances at 6 -29.1 and 37.2, respectively.A sample was contained in a thin-walled, 16mm diameter vanadium sample can. A neutron diffraction pattern was measured at 363 K in the furnace on the instpment D2B at ILL, Grenoble, with a wavelength of 1.5939A. Prior to the scan the sample was heated in situ to ensure it was still dehydrated. This sample was from the same batch as that which was used for the synchrotron study. For the X-ray study, the sample was loaded into a flat plate sample holder in a glove box. The data were collected at room temperature under a dry nitrogen atmosphere using a glove box which surrounded the goniometer head on station 8.3 of the Daresbury synchrv- tron radiation source, using a wavelength of 1.4508 A. Compared to the previously published structures, the diffrac- tion pattern showe! splitting of some peaks, e.g.the peak at a d-spacing of 4.508 A (at 28 =18.55") which would correspond to the (210) reflection in the hexagonal cell of the as-synthesised J. Muter. Chem., 1996,6(11), 1831-1835 1831 material. These splittings indicate that there is a distortion of the hexagonal cell when the sample is calcined, resulting in the loss of the six-fold symmetry axis. A C-centred orthorhombic cell, with twice the volume of the hexagonal cell, can be derived by applying the transformation: a' =a; b' =a +2b; c' =c (see International Tables of Crystallography, 1987, pp. 78-79). The highest symmetry space group for this new cell is Ccc2 (no. 37), which is a direct sub-group of P6cc (no.184), retaining all symmetry elements except for the six-fold axis. Three times as many independent atoms are required to describe the structure in this space group. Some additional weak peaks are also visible in the pattern, e.g. the peaks at 28= 18.90 and 19.56 in Fig. 1. These may indicate that the symmetry is lower than Ccc2. For example, the weak peak at 18.90 could appear if the unit cell is not C-centred, but primitive. However, most of the new peaks cannot be indexed on the basis of a primitive cell and are probably due to minor impurities introduced by breakdown of the ALP04-5 during calcining, or perhaps indicate a more complex superstructure, which would be beyond the scope of the current powder measurements.To check the suitability of lower-symmetry variants of the Ccc2 cell, all the space groups directly derived from Ccc2 need to be investigated: orthorhombic Pcc2 (no. 27), Pcn2 (no. 30), Pnc2 (no. 30) and Pnn2 (no. 34); monoclinic C112 (P2, no. 3), Clcl (no. 9) and Ccll (no. 9). Recently the lattice parameters of an as-synthesised ALP0,- 5, which uses a synthesis medium containing fluoride ions and tropine as a template, have been reported in the orthorhombic spaceo group Ccc2 with cell parameters a = 13.78, b =23.33, c = 8.44 A;I4 a reasonable fit between observed and calculated X-ray diffraction patterns was reported using the atomic coordi- nates derived from one of the earlier papers, but the authors did not pursue the matter further.For the current inv!stigation a similar cell, with a= 13.76, b =23.76 and c= 8.38 A derived from the hexagonal cell of Qiu et al.,' was checked against the synchrotron data by using the profile decomposition procedure of LeBail et a1.15 using the program MPROFIL16 and con- I 1 14400'"I *"I' I O' 16.5 19.0 19.5 20.0 28/degrees Fig. 1 Part of the synchrotron pattern of calcined ALP04-5, illustrat-ing the splitting of the (210) reflection at 18.52' confirming the lowering of symmetry from hexagonal to orthorhombic, and additional weak peaks at 28= 18.90 and 19.56" verged to R,,=14.3% and R,,,=8.1%. With the profile parameters obtained from this refinement and atomic positions generated by transforming the symmetry from the structure of Qiu et al.,' Riet~eld'~9~'refinements were performed in the different space groups.The R, values and cell parameters for the last cycle of refinements using the synchrotron data are presented in Table 1. The restraints which were applie! during the refinements were: oA1-O distances 1.71 f0.005 A, P-0 distances 1.51 k0.005 A and the tetrahedral angles around the A1 and P 109.5 f.3". Whereas a joint refinement from the synchrotron and neu- tron data is desirable, it would not be wise to attempt one with a temperature difference of nearly 100 K between the two measurements. Therefore, separate refinements of the neutron data were carried out, using as starting models the coordinates derived from the synchrotron refinements. The resolution of D2B (ILL, Grenoble) is lower than that of station 8.3 (Daresbury) and no splitting of peaks was apparent in the diffraction pattern.On the other hand, the greater contrast between the neutron scattering lengths of A1 and P (3.449 and 5.130 fm respectively) compared to their X-ray scattering factors provides a rigorous test of the models. It was again necessary to employ restraints to achieve convergence, but they were applied less stringently than when refining from the synchrotron data. In the final stages of the refinemfnts, the restraints applied weFe: A1-0 distances 1.71 k0.05 A, P-0 distances 1.51f.0.05 A and the tetrahedral angles around the A1 and P 109.5+5". Several extra weak peaks were detected in this pattern which are due to titanium, which forms part of the endcaps of the sample can.The final cell parameters, R values and ranges and average values of the bond distances and bond angles for the eight models tested are shown in Table 2. Of the several models tested, the one that fits both the synchrotron and the neutron data best is the model in Pcc2. The model does not account particularly well for the weak peaks, some of which are partially fitted, e.g. the peak at 18.90". However, these peaks are very weak and their contribution to the model is slight, and it seems likely that there is at least one (unidentified) impurity phase. The key difference between the model in Pcc2 and the other models is that there is an increase in the flexibility of the structure due to the lowering of symmetry, which gives rise to two crystallographically distinct channels (see Fig.4 later). This less rigid structure has more consistent and chemically correct bond distances and angles, if compared, for example, with the more symmetrical model in Ccc2 where unreasonable tetrahedral angles of 77.5" were obtained (see Table 2). A similar situation has also been seen recently in a combined synchrotron and neutron powder diffraction study on the zeolite ferrierite where an unfeasible T-0-T angle of 180" was eliminated when the crystallo- graphic symmetry was reduced.lg Details of the refinement in space group Pcc2, which appears to be the most satisfactory, are shown in Table 3. A full discussion will be confined to this space group.A plot of the observed, calculated and difference profiles for the synchrotron Table 1 Comparison of X-ray powder profile refinements using synchrotron radiation in various subgroups of P6cc space group a/A b/A CIA aldegrees Pldegrees yldegrees Rw, (%I c112 Clcl Ccll ccc2 Pcc2 Pcn2 Pnc2 Pnn2 13.7203 (8) 13.7 176( 6) 13.7 177 (7) 13.7591(6) 13.7593(6) 13.7571 (7) 13.7595( 6) 13.7170(6) 23.829( 1) 23.832( 1) 23.832( 1) 23.759( 1) 23.758( 1) 23.760( 1) 23.759,(1) 23.832( 6) 8.3802( 3) 8.3801 (3) 8.3 804( 3) 8.3800( 3) 8.3801(3) 8.3797( 3) 8.3801 (3) 8.3801 (3) 90 90 90.01(3) 90 90 90 90 90 90 90.02( 2) 90 90 90 90 90 90 90.09( 1) 90 90 90 90 90 90 90 18.4 18.4 18.8 17.8 17.8 18.0 18.0 18.4 1832 J. Muter.Chem., 1996, 6(11), 1831-1835 Table 2 Comparison of neutron powder profile refinements in subgroups of P6cc Shortest, longest and average bond distances and angles space group a/A b/A CIA aldegrees Pfdegrees yldegrees R,, (%) x2 Al-O/A P -O/A 0-A1-Oldegrees 0-P-Ofdegrees Al-0-Pfdegrees c112 13.798 (4) 23.900( 8) 8.4181 (6) 90 90 90.00(4) 16.5 7.3 1.6489 1.4677 92.69 88.87 143.48 1.7378 1.5662 117.47 120.39 174.56 1.71(2) 1.51(2) 109(4) 109(6) 152(7) Clcl 13.798(4) 23.904( 7) 8.4182( 6) 90 90.13(2) 90 15.9 6.6 1.6449 1.4134 98.10 93.61 146.53 1.7428 1.6187 135.34 125.64 169.54 1.70(2) 1.52(4) 109(6) 109(6) 152(5) CCl1 13.799 (4) 23.906( 6) 8.4195(6) 89.88(2) 90 90 16.4 6.8 1.6618 1.4380 77.50 96.34 140.42 1.7694 1.6179 120.02 122.73 170.57 I.70 (3) 1.52(5) 109(7) 109(6) 151(6) Ccc2 13.799( 3) 23.904( 5) 8.4188(6) 16.8 6.9 1.6509 1.4903 95.73 93.16 149.30 1.7340 1.5425 119.55 122.68 160.90 I.70(2) 1.52(2) 109(5) 109(6) 152(3) Pcc2 13.794( 3) 23.900( 6) 8.4168( 6) 15.8 6.5 1.6748 1.4424 101.32 99.22 143.88 1.7527 1.5616 119.23 116.89 170.79 1.71(2) 1.51(3) 109 (4) 109(4) 152(6) 4 Pcn2 13.786(1) 23.932(2) 8.4199( 6) 24.5 16.7 1.6293 1.4548 104.94 104.25 143.13 1.7254 1.5581 115.99 116.08 176.54i5 1.69(2) 1.53(2) 109(3) 109(3) 152(9)rc Pnc2 13.787(1) 23.935(2) 8.4215(7) 23.6 15.0 1.6439 1.4745 103.82 104.96 138.24 r 1.7083 1.5930 114.60 113.92 167.66Q 1.68(2) 1.54 (3) llO(3) 109(3) 151(7) m Pnn2 13.788( 1) 23.927( 2) 8.4210( 6) 20.6 11.4 1.6625 1.5050 99.78 102.47 141.93J" 1.7338 1.5788 115.95 118.00 175.20 1.69(2) 1.53 (2) 109(3) 109(4) 151(7) w 00 w+ I w 00 wul Y Table 3 Crystallographic data for the neutron Rietveld refinement of ALPO4-5 molecular formula W,P)O,731.712MY a/-$ 13.794( 3) b/-$ 23.900( 6) c/A 8.4 168 (6) space group Pcc2 (no.27) V/P 2775(3)z 4 DJg cm-3 1.751 T/K 363 WA 1.5939 R, (Yo)" 15.8 x2 6.5 number of data points 2537 number of contributing reflections 2852 degree of freedom (N -P +C) 2418 'Rwp=loo(Cwj(yi -y,j)2/Zwiy:} I". bx2=(CW, (yi -Y,~)~/(N-P +C)}. N is the number of contributing observations, P is the number of parameters, C is the number of restraints. wi=l/y(total) =l/yi + background. 32400 -4 20 30 40 50 60 70 80 26Vdegrees Fig.2 Observed (*),calculated (-) and difference plot of the final refinement of calcined ALP0,-5 from synchrotron X-ray powder diffraction data. Reflection positions are shown as vertical lines (note the square-root scale). 0-Ill I I IIIIII I Ill I IIIIr Int IIIIIIYIIIIIIIII-20 40 60 80 100 120 14028/degrees Fig. 3 Observed (*),calculated (-) and difference plot for the final refinement of calcined ALP0,-5 from neutron powder diffraction data. Reflection positions are shown as vertical lines from ALP04-5 (lower) and spurious titanium endcaps (upper). and neutron data are shown in Fig. 2 and 3. A diagram of the structure is illustrated in Fig. 4. The overall structure of the material is compatible with that previously reported.*-'' The two distinct one-dimensional 12- rings channels run parallel to c and are almost circular, with trans-ring O...O distances of 10.01-10.13 A.This feature can be contrasted to that in the as-synthesised form where a slight distortion in the channel was observed. 1834 J. Muter. Chem., 1996, 6(ll), 1831-1835 e C Fig. 4 Diagram of the structure of calcined ALP0,-5 as derived from the neutron powder diffraction Rietveld refinement The final atomic parameters from the neutron refinement are shoyn in Table 4. The ayerage values of the A1-0 [1.71(2) A] and P-0 [1.51(3) A] distances are in fair agreF- ment with those found in berlinite2' [1.736( 1) and 1.521 (1)A, respectively]. The average Al- 0-P angles [152( 6)"] are also in accord with those found in microporous ALPOs.It is also interesting to note that the refinement from the neutron data is more robust. This is due to the higher quality of the neutron data out to high sinO/A and because of the reduction in the apparent pseudo-symmetry in the diffraction pattern caused by the enhanced contrast between P and Al. However, without the superior resolution of the synchrotron data the reduction of symmetry from hexagonal to orthorhombic would not have been so obvious and hence it would not have been possible to obtain the correct starting model for the refinement. This highlights the advantages of using both sets of data in a complementary manner. Atomic coordinates, thermal parameters, and bond lengths and angles have been deposited at the Cambridge Crystallographic Data Centre (CCDC).See Information for Authors, J. Muter. Chem., 1996, Issue 1. Any request to the CCDC for this material should quote the full literature citation and the reference number 1145/11. In conclusion, we have shown that in our sample at tempera- tures between room temperature and 363 K the structure of ALP0,-5 is best described in the orthorhombic space group Pcc2. This permits a strictly alternating, fully ordered structure with reasonable bond lengths. Table 4 Atomic coordinates for calcined AlPO-5 determined from References Rietveld fit to neutron diffraction data 1 S. T. Wilson, B. M. Lok, C. A. Messina, T. R. Cannan and atom X Y Z Biso 2 E.M. Flannigen, J. Am. Chem. Soc., 1982,104,1146. S. T. Wilson, B. M. Lok and E. M. Flannigen, US Pat., 4310440, 0.294( 2) 0.292( 4) 0.315(5) 0.3 13( 4) 0.191(3) 0.363( 5) -0.389( 3) -0.389( 4) -0.319( 5) -0.372(4) -0.365(6) -0.492( 4) 0.107( 3) 0.111( 3) 0.003 (4) 0.12 1 (5) 0.174( 4) 0.129(5) 0.789( 3) 0.788(4) 0.81 l(2) 0.8 10( 4) 0.684(4) 0.853(3) 0.107 (3) 0.104( 4) 0.182( 5) 0.098(4) 0.136( 6) 0.012(4) 0.605( 3) 0.608(4) 0.504 (4) 0.622(4) 0.675( 5) 0.633(5) 0.169 (2) 0.169(2) 0.108(2) 0.177( 2) 0.181 (3) 0.206( 2) 0.063 (2) 0.061 (2) 0.103( 3) 0.069( 2) 0.003( 2) 0.078( 3) -0.226( 2) -0.226( 2) -0.211( 3) -0.217( 3) -0.184( 3) -0.284( 3) 0.665 (2) 0.667(2) 0.608 (2) 0.660( 2) 0.682( 3) 0.708( 1) 0.560( 2) 0.561 (2) 0.604( 3) 0.559( 3) 0.504( 2) 0.578( 3) 0.273 (2) 0.274( 2) 0.288( 3) 0.288(2) 0.314(3) 0.214( 2) 0.130 0.497( 6) 0.102( 8) 0.297(5) 0.075 (7) 0.050(7) 0.136( 7) 0.5 17(7) 0.056( 8) 0.3 16( 8) 0.09(1) 0.07(1) 0.122( 7) 0.500(6) 0.074( 9) 0.297( 7) 0.040(8) 0.054( 8) 0.13 0.503( 5) 0.049( 7) 0.304(5) 0.109(1) 0.050(7) 0.112( 7) 0.484( 7) 0.091 (8) 0.283 (7) 0.044( 9) 0.034( 9) 0.141(7) 0.5 12 (6) 0.083 (9) 0.3 12( 7) 0.069( 8) 0.100( 8) 2.5(3) 1.5(4) 3.64(8) 3.64(8) 3.64(8) 3.64(8) 2.5(3) 1.5(4) 3.64(8) 3.64(8) 3.64( 8) 3.64(8) 2.5(3) 1.5(4) 3.64(8) 3.64(8) 3.64(8) 3.64(8) 2.5(3) 1.5(4) 3.64(8) 3.64(8) 3.64(8) 3.64(8) 2.5(3) 1.5(4) 3.64(8) 3.64(8) 3.64(8) 3.64(8) 2.5(3) 1.5(4) 3.64( 8) 3.64( 8) 3.64(8) 3.64(8) 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 1982.J. J. Pluth, J. V. Smith and J.M. Bennett, Acta Crystallogr., Sect. C, 1986,42,283. B. M. Lok, C. A. Messina, R. L. Patton, R. T. Gajek, T. T. Cannan and E. M. Flannigen, J. Am. Chem. SOC.,1984,106,6092. P. A. Wright, R. H. Jones, S. Natarajan, R. G. Bell, J. Chen, M. B. Hursthouse and J. M. Thomas, J. Chem. SOC., Chem. Commun., 1993,633. R. M. Dessau, J. L. Schenker and J. B. Higgins, Zeolites, 1992, 12, 13. M. E. Davies, C. Saldarriage, C. Montes, J. Garces and C. Crowder, Nature (London), 1988,331,698. J. M. Bennet, J. P. Cohen, E. M. Flangen, J. J. Pluth and J. V. Smith, ACS Symp. Seu., 1983,218, 109. S. Qiu, W. Pang, H. Kessler and J. L. Guth, Zeolites, 1989,8,440. J. W. Richardson, J. J. Pluth and J. V. Smith, Acta Crystallogr., Sect. C, 1987,43, 1469. A. J. M. de Man, W. P. J. H. Jacobs, J.P. Gilson and R. A. van Santen, Zeolites, 1992, 12, 826. D. Muller, E. Jahn, B. Fahlke, A. Ladwig and U. Haubenreisser, Zeolites, 1985,5, 53. M. P. J. Peeters, L. J. M. van de Ven, J. V. de Haan and J. H. C. van Hooff, J. Phys. Chem., 1993,97,8254. N. Ohnishi, S. Qiu, 0. Terasaki, T. Kajitani and K. Hiraga, Microporous Muter., 1992,2,73. A. LeBail, H. Duroy and J. L. Fourquet, Muter. Res. Bull., 1988, 23,447. A. Jouanneaux, A. D. Murray, A. N. Fitch, MPROFIL: a program for LeBail decomposition of powder patterns, 1990. J. K. Cockcroft, PROFIL: A Rietveld program for the refinement of crystal structures from single and multi-phase powder neutron or synchrotron radiation data, Institut Laue Langevin, Grenoble, France, 199 1. H. M. Rietveld, J. Appl. Crystallogr., 1969,2,65. R. E. Morris, S. J. Weigel, N. J. Henson, L. M. Bull, M. T. Janicke, B. F. Chmelka and A. K. Cheetham, J. Am. Chem. SOC., 1994, 116,11849. 20 N. Thong and D. Schwanzenbach, Acta Crystallogr., Sect. A, 1979, We thank Dr. Dewi D. Lewis for useful discussions concerning 35, 658. their recent theoretical calculations on ALPO,-5.? A. J. M. 21 A. R. Ruiz-Salvador, G. Sastre, D. W. Lewis and C. R. A. Catlow, would like to acknowledge Universidad de 10s Andes and following paper. CONICIT-Venezuela for financial support. Paper 6/01875K; Received 18th March, 1996 7 Note added at proof: A recent theoretical calculation21 has shown that the proposed lowering in symmetry is energetically favourable. J. Muter. Chew., 1996, 6(ll), 1831-1835 1835

ISSN:0959-9428    

DOI:10.1039/JM9960601831

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Space group symmetry and Al- 0-P bond angles in A1P04-5 A. Rabdel Ruiz-Salvador,+ German Sastre, Dewi W. Lewis and C. Richard A. Catlow" Davy Faraday Laboratory, Royal Institution of Great Britain, 21 Albemarle St., London, UK W1X 4BS Static lattice simulations are presented on the microporous aluminophosphate A1PO4-5. Minimisation methods together with lattice dynamical calculations are able to identify stable minima for the structure. We show, in agreement with the recent work of Henson et al., that reduction of the symmetry of the structure from P6cc to P6 leads to the relaxation of A1-0-P angles from the linear structure reported in earlier crystallographic structures. We find that the orthorhombic space group Pcc2 suggested recently by Mora et ul. has only a very slightly higher energy than that calculated for the P6 structure. Differences in bond lengths and angles between the two structures are correspondingly small.We find that the inclusion of a representation of the polarisability of the oxygen in the potential model is essential in removing the linear A1-0-P angles in the simulated structures. Since their discovery in the early 1980s,' the aluminophosphate (AlPO) family of molecular sieves has been studied extensively as their structural and chemical properties are analogous to zeolites and applications in ion exchange and catalysis have been realised. Many AlPO structures are isostructural with those previously found as aluminosilicates, for example A1P04-17 (erionite),' SAPO-34 (~habazite)~ and SAPO-37 (fa~jasite).~However, others have unique structures found only as AIPOs, for example DAF-15 and VPI-5.6 The A1P04-5 structure was also a unique AlPO structure until the sub- sequent synthesis of SSZ-24, its siliceous analog~e.~ A1P04-5 was one of the first structures of the alumino- phosphate family to be synthesised and, given its unique structure, it was the focus of several crystallographic studies.Although the structure was determined from single-crystal X-ray diffraction studies'.' a number of important questions remain regarding detailed features of the structure. The mate- rial consists of 4, 6 and 12-membered rings which form a hexagonal arrangement of parallel channels, bounded by 12- membered rings, parallel to the [OOl] direction (Fig.1). The initial crystallographic were performed on as-prepared materials (containing the template, tetrapropyl- ammonium) and refined to the hexagonal space group P6cc. They give a good description of the channel structure, but one of the oxygen sites has a rather high temperature factor and the Al-0-P angles involving this oxygen are close to 180". One set of authors' attribute this observation to static disorder of the oxygen which is structurally distributed about three equivalent sites around the Al-P axis as shown in Fig. 2. In their neutron diffraction study of the calcined material, Richardson et u1." found that a higher hexagonal symmetry of P6/mrnc was required to obtain a satisfactory refinement.However, this model requires the A1 and P to be disordered over the tetrahedral sites, which is contrary to the strict alternation of A1 and P found in these materials by other methods. Furthermore, a low-temperature pattern obtained by these authors proved impossible to refine in a hexagonal space group. NMR results also suggest that the symmetry was lower than that reported from crystallographic studies." Theoretical studies have in general found the same structural properties as determined e~perimentally,'~-'~ although in general these calculations have suggested that lower symmetries are possible. The recent paper by Henson et appears to confirm the presence of a lower symmetry with their calculations demon- strating that a hexagonal unit cell with space group P6 is not t Permanent address: Zeolite Engineering Laboratory, Department of Materials Science, IMRE-Faculty of Physics, University of Havana, 10400 Havana, Cuba.only more energetically favoured (by 0.92 kJ mol-l) but also contains no linear Al-0-P angles. Furthermore, these authors find that the inclusion of a representation of the polarisability of oxygen is crucial in lowering the symmetry of such systems. We note that other potential models, which do not include polarisable oxygen ions, do not remove the linear Al-0-P angles in A1PO4-5, although a lower symmetry is suggested by these calc~lations.~~ Recent structure determinations have also alluded to alterna- tives to the P6cc space group. A reversible phase transition from hexagonal to orthorhombic was reported for an MPo4-5 material in the presence of the template molecule, tropine.16 Furthermore, the recent work of Mora et also finds that the structure, this time of a calcined material, is best refined Fig.1 Topology of A1P04-5 showing 12-membered ring channels and hexagonal arrangement. Unit cell shown is that of Bennett et a1.8 Fig. 2 Disorder around the Al-P axis of oxygen. The average oxygen position results in the linear Al-0-P configuration found in the diffraction studies. J. Muter. Chem., 1996,6( ll), 1837-1842 1837 in the orthorhombic space group Pcc2. Their room-tempera- ture high-resolution powder X-ray diffraction pattern shows visible splitting of the peaks corresponding to the (210) reflec- tion in the hexagonal cell, confirming the need to refine in a lower symmetry. Although this splitting is not evident in their neutron diffraction data, these data can be successfully refined in the same unit cell.The lowering of the space group has the effect of removing any linear A1-0-P angles. Linear T-0-T angles are also obtained in refinements of other microporous crystal structures. However, it appears that they occur as a consequence of the diffraction refinement being performed in a space group which has too high symmetry. For example, the structure of ferrierite was originally refined in Immm,18 but subsequently, using a combination of high- resolution neutron diffraction and computer modelling, a structure with the lower space group of Pmnn (and no linear angles) was determined by Morris et uZ.l9 In the light of the questions raised here, regarding both constraints imposed by the refinement process and the effect on simulated structures of different interatomic potentials, we have applied lattice energy minimisation methods to investigate the structure of A1PO4-5.Specifically, we have addressed whether (i) a hexagonal unit cell can describe the structure without recourse to either random Al/P distributions or linear Al-0-P angles; (ii) the orthorhombic cell of Mora et uZ.l7 is a viable alternative; and (iii) different potential models can correctly model Al-0-P angles. The recent paper by Henson et ~1.l~showed that a space group of P6 removed the linear angles.Here we will consider the relationship between such a hexagonal cell and other related space groups in terms of the structural parameters of the material. Methodology The calculations in this work were performed using lattice energy minimisation techniques2’ with the GULP code21 being employed throughout. The methodology uses standard tech- niques based on the Ewald22 method for summation of the long-range Coulombic interactions, and direct summation of the short-range interactions which can be described by a number of standard formulations. All atomic coordinates and cell parameters are optimised to zero force using the Broyden- Fletcher-GoldfarbShanno (BFGS)23 and rational function optimisation (RFO)24 minimisation methods.A convergence criterion of a gradient norm below 0.001 eV A-‘ was used for all the calculations. These techniques have been successful not only in modelling the structure of microporous material^^^.^^ but also have been employed as an aid to solving the structure of such rnaterial~.~~*~~,~~ A number of calculations have been carried out, not only with the single (both hexagonal8 and orthorh~mbic’~) A1P04-5 unit cell but also with several supercells in order to remove any constraints imposed by the single unit-cell calcu- lations, which will be discussed below. Analyses of the phonon frequencies were performed in the optimised structures in order to check whether a true minimum had been located; imaginary frequencies indicate that the minimisation procedure has located a saddlepoint (and not a minimum) on the potential- energy surface.Although the BFGS23 minimisation algorithm is very efficient at locating minima, it can, occasionally, converge to a saddle-point if the energy surface is very shallow. Such a saddle-point is manifested as imaginary vibrational frequencies in subsequent lattice phonon calculations. The RF024 algor- ithm, on the other hand, conditions the second derivative (Hessian) matrix to be of the correct form for a true minimum; that is, with no negative eigenvalues. Such a minimisation scheme ensures convergence to a true minimum. However, this is at the expense of additional computational cost and thus this minimisation procedure is only used when the BFGS does not successfully find a minimum. Two sets of interatomic potentials were used to model the interactions in the structure.The first set, by Gale and Hen~on,~~included the following terms: Coulombic interaction with formal charges on the ions, short-range pair potentials (described by a Buckingham function) and a three-body, bond- bending term. The shell model3’ was used to simulate the polarisability of the oxygen atoms. This set of potentials was parameterised to reproduce the structure of berli~~ite~~ and has been shown to model successfully a number of AlPO structures including AlPO4-5.l5 The second set of potentials, by van Beest et uses a partial charge model and a (Buckingham) two- body short-range interaction potential, with no treatment of polarisability or bond-bending terms.The parameters were derived from quantum mechanical calculations on small clus- ters and have been applied to modelling a number of AlPO systems.13 All the potential parameters used are given in Table 1, which also lists the explicit form of the potential used. Further details of the methodologies employed2’ and their application to the study of microporous solids32 are available elsewhere. Table 1 Potential parameters used in the GULPz1 code* (a) Gale and Henson potential^:^' Buckingham potentials A/eV PlA C/eV A6 2A13+ -0-1283.90 0.32052 10.66 2p5+-0-877.34 0.35940 0.00 02--02-22764.00 0.14900 27.88 three-body potential k/eV rad-’ O,/degrees O-T-Ob 2.09724 109.47 core-shell potential k/eV A-2 O2-74.92 Coulombic +charges P5 ~13 02-(core) 02-(shell)+ +5 +3 +0.86902 -2.86902 (b)van Beest et UZ.:~~ Buckingham potentials A/eV PlA c/ev A6 2A13-t-0-16008.53 0.20848 130.56 2p5+-0-9034.21 0.19264 19.88 02--02-1388.77 0.36232 175.00 Coulombic +charges P5 AP + O2-3.4 1.4 -1.2 “Two sets of parameters are reported; those of Gale and Hern~on~~ and those of van Beest et aL3’ A cutoff value of 16 A was used for the short-range Buckingham potentials. These parameters are employed in the general energy expression given by: V=ZKj(Buckingham)+ CKj(Coulombic)+CJ(jk(threebody)+ ZKj(core shell) where, Kj(Buckingham)=Aij exp(-rij/p)-Cij/r6 Kj(Coulombic)=(qiqj)/rij Kjk(three body) =0.5kij,(Oijk -Ooijk)2K (core shell)=0.5kiAr? where Ar is the core-shell separation.b~=~ior P. 1838 J. Muter. Chem., 1996,6(11), 1837-1842 Results and Discussion The starting point for these calculations was the hexagonal P6cc structure of A1P04-5,' having one equivalent T (A1,P) position and four different A1-0-P angles (Fig. 3) designated Al-O(i)-P (i= 1-4). In the reported crystal structure,' three of these angles have values between 145 and 155" while the fourth [Al-O(2)-P] is reported to be nearly 180". Previo~slyl~-~~it has been found that energy minimisation of this unit cell results in a structure which is in good agreement with the experimental data, and indeed, retains the almost linear A1-O( 2)-P bonds; these earlier results are reproduced in Table 2.However, subsequent lattice dynamical calculations on our previous minimum structure,14 using the parameters of Gale and Hens01-1,~~ have revealed that this optimisation does not produce a true minimum-energy structure, as noted recently by Henson et ul." and previously by de Man et ~1.'~ Attempts to remove the imaginary frequencies whilst retaining the constraint of the P6cc space group failed, suggesting that a reduction of symmetry was required. An analysis of the eigenvectors of the imaginary frequencies generated show that the major components are on the oxygen atoms which form the linear Al-0-P bridge, indicating again that there is strain in the position of these oxygens.We have therefore introduced a number of modifications to the unit-cell model in order to investigate further the space group of A1P04-5. In particular, we examine the possibility of disorder in the position of the O(2) atom as suggested in the original structure determination' and as seen in the orthorhombic cell of Mora et u1.17 The A1P04-5 channel structure is formed by 12T (A1,P) atoms parallel to the crystallographic direction [OOl] (Fig. 1). The arrangement of the TO4 units in this structure is such that the anomalous Al-O(2)-P angles are parallel to the [OOl] direction, and hence parallel to the channel system (Fig. 3). The position of the O(2) atom is constrained by the space-group symmetry and therefore full optimisation of the coordinates of this atom requires the removal of such symmetry constraints.In order to determine if this constraint can be removed, we have performed further calculations using two Fig. 3 Hexagonal P6cc unit cell of Bennett et aL,* showing the linear Al-O(2)-P angle different approaches. First, we consider a P1 cell derived from the hexagonal cell, but, with all the O(2) atoms randomly displaced from their original sites. Secondly, we consider supercells of the original hexagonal cell. All the following calculations were performed using the Gale and Henson poten- tial Taking the experimental P6cc structure, we have re-optimised (using the BFGS algorithm) the structure of the single unit cell but with the coordinates of the O(2) oxygens randomly displaced from their original symmetric positions.This method allows the Al-O(2)-P angles to adopt a value different from 180". The final optimised structure, shown in Fig. 4, is very similar to the initial one but two important differences should be stressed. First, although the crystallo- graphic parameters remain hexagonal (Table 3), the symmetry has decreased from P6cc to P6. Secondly, the Al-O(2)-P angles now have values ranging from 155 to 160" with all the Al-0-P angles being in the range 145-160". The energy of this unit cell (Table 3) is lower than that of the unit cell which retains the linear Al-O(2)-P angles.14 An analysis of the vibrational modes of the optimised structure showed no imaginary frequencies demonstrating that this structure corre- sponds to a true minimum.This optimised cell is the same as the P6 cell found by Henson et aZ.15 We now constructed supercells of the initial P6cc structure and optimised them using the BFGS algorithm. The supercells used were 1x 2 x 1, 1 x 1 x 2 and 2 x 1 x 2 with respect to the initial cell in the directions [1001, [OlO] and [OOl]. It should be noted that when no replication of the cell in the c direction is performed, as in the case of the 1 x 2 x 1 supercell, no differences should be expected with this technique with respect to the results obtained for the 1 x 1 x 1 cell since the constraint on the Al-O(2)-P angle remains. The final results obtained in these three optimisations are shown in Table 4.We observe that, as expected, the results for the 1 x 2 x 1 cell are not significantly different from those obtained for the primitive (1 x 1 x 1) unit cell (Table 2). On the other hand, BFGS optimisation of both the 1x 1 x 2 and 2 x 1 x 2 supercells allows the optimisation of some of the linear Al-O(2)-P to ___ -I-------I r I Ir-rr--r-I__-Fig. 4 Comparison of the structures of the P6 (top) and P6cc (bottom) unit cells obtained in this work Table 2 Optimisation" of AlPO,-5 hexagonal P6cc cell from de Man et all3 cell a/A CIA A1-0-P/degrees E(A1POJb/eV crystallographically refined structure' 13.771 8.379 146.6-176.1 - Gale and Henson potentials2' BFGS 14.183 8.678 153.9-178.4 -268.0408 (8 imaginary modes) van Beest et al.(no 13.979' 8.425 144.2-178.7 -117.9448 imaginary modes) "Using the potentials of van Beest et aL31 and from Sastre et (using the potentials of Gale and Hen~on~~). Note that the lattice energies are not comparable owing to the difference in potential parameters (most notably the charges). bEnergy per AlPO, unit. 'Average value: a= 13.8201 and b = 14.1375. J. Muter. Chem., 1996,6(11), 1837-1842 1839 Table 3 Optimised geometry of a P1 cell derived from the hexagonal cell, using the potentials of Gale and Hen~on.~' The O(2) atoms were all randomly displaced from their experimental sites to break the symmetry constraint of the space group optimisation mode a/A CIA afdegrees Pfdegrees yfdegrees Al-0-Pfdegrees E(AlPO,)"/eV BFGS 13.831 8.547 89.979 89.992 119.982 147.7- 177.0 -268.0448 (9 imaginary modes) RFOb 13.759 8.393 90.000 90.000 120.000 141.8-154.7 -268.0643 (no imaginary modes) "Energy per AlPO, unit.bThe space group of the final optimised cell is P6. Table 4 Supercells of the hexagonal unit cell all optimised using the BFGS method. The potentials of Gale and Henson2' were used" supercell a/A bfA CIA afdegrees Pfdegrees yfdegrees A1-0-P/degrees E(AIPO,)bfeV 1X2X1 13.831 27.662 8.547 89.977 89.993 119.983 147.3-176.9 -268.0448 1x 1x2 13.828 13.828 16.875 90.000 90.000 119.999 146.7-177.1 -268.0369 2 x 1x2 27.539 13.787 16.870 90.019 90.141 119.881 143.4- 171.5 -268.0479 "All cells exhibit imaginary phonon frequencies after BFGS minimisation, see Table 5 for final optimised configurations.bEnergy per AlPO, unit. Table 5 Supercells of the hexagonal unit cell all optimised using the RFO method (all imaginary frequencies are removed). The potentials of Gale and HensonZ9 were used supercell a/A bfA CIA afdegrees Pfdegrees yldegrees Al-0-Pfdegrees E(AlPO,)"/eV 1X2X1 13.768 27.519 8.405 90.030 89.906 119.908 142.5-161.6 -268.0612 1x 1x2 13.761 13.761 16.810 90.000 90.000 120.000 142.4-1 61.5 -268.0541 2 x 1x 2 27.521 13.761 16.810 90.000 90.000 120.000 144.0-1 56.7 -268.0541 "Energy per AlPO, unit more reasonable values. However, the distribution of A1-0-P angles remains similar to those found in the BFGS minimisation of the primitive cell, a number of the Al-O(2)-P angles remaining near 180" and the final optimised structures retain a number of imaginary vibrational frequencies. An interesting point at this stage is that the 2 x 1 x 2 supercell is not hexagonal.An analysis of the sym- metry of this optimised structure showed that it corresponds in fact to an orthorhombic group, having crystallographic parameters very similar to that reported by Mora et and with a unit cell containing the same number of atoms. However, as before, phonon calculations on the three optimised supercells reveal imaginary vibrational modes. Further optimisations using the RFO algorithm were performed to remove these imaginary vibrational modes and the final minimised unit cells Fig.5 Unit cell of the optimised P6 cell showing the notation used are shown in Table 5. Comparison to the previous structures in Table 6 with imaginary modes (Table 4) shows that they are all lower in energy, as expected.In all the three cases the Al-O(2)-P aluminophosphates in general. We conclude therefore that the angles are no longer 180". The symmetry of the cells remain hexagonal space group obtained is not a result of artificial hexagonal but is reduced to P6 and furthermore the 2 x 1x 2 constraints.supercell now reverts to a hexagonal cell. The analysis of the We now consider the orthorhombic Pcc2 cell determined by bond lengths and bond angles for this P6 cell is shown in Mora et again performing BFGS minimisation followed Table 6 and the notation used is described in Fig. 5. The values by RFO optimisation to remove the imaginary vibrational obtained compare well with all previous experimental values modes. As for the hexagonal cell, the optimised unit-cellfor A1P04-5,8-10916917 and are also typical of those found in parameters, bond lengths and angles are all in good agreement with the experimental results (Table 7).Furthermore, the Table 6 Analysis of the bond angles and bond lengths in the calculated orthorhombic cell is also lower in energy than the hexagonal P6 cell. Atom labels correspond to those in Fig. 5 (P6)cell before removal of the imaginary vibrational frequen- cies. However, the energy of the cell is not as low as that found distancefA anglefdegrees in the fully optimised hexagonal phase. We would expect lower symmetry cells at lower temperatures, but, we note that the P-0 A1-0 P-0-A1 experimental data17 were collected at room temperature. However, the differences are small and the change in lattice 1.502 1.716 149.06 1.519 1.712 154.19 energy (cu.0.5 kJ mol-I per AlP04 unit) could be outweighed 1.520 1.734 141.95 by an entropy term. Furthermore, the differences between the 1.528 1.724 146.71 hexagonal and orthorhombic cells do not reflect any substan- 1.507 1.729 145.25 tial changes in bond lengths and bond angles. This is illustrated 1.512 1.708 154.43 in Fig. 6 which shows both the optimised P6 and Pcc2. The 1.523 1.724 146.53 extremely small differences in lattice energy show how these 1.530 1.731 143.49 subtle modifications in symmetry can be brought about with 1840 J.Muter. Chem., 1996,6( ll), 1837-1842 Table 7 Calculated unit-cell parameters for the orthorhombic Pcc2 unit cell of Mora et a1.l' using the parameters of Gale and Hen~on~~ cell a/A b/A c/A aldegrees Pldegrees yldegrees Al-0-Pldegrees E(A1PO4)a/eV e~perimental'~ 13.797 23.899 8.417 90.000 90.000 90.000 147- 156 -BFGS 13.768 23.963 8.417 90.000 90.004 89.978 141.5-166.3 -268.0531 (8 imaginary modes) RFO 13.754 23.900 8.417 89.999 89.999 90.398 140.1-156.5 -268.0587 (no imaginary modes) "Energy per AlPO, unit. Table8 Optimised geometry" of cells derived from the hexagonal P~cc,' P6 (this work) and orthorhombic Pcc2I7 cells using the potentials of van Beest et aL31 cell a/A b/A CIA aldegrees flldegrees yldegrees Al-0-Pldegrees E(AIPO,)b/eV P6cc 14.172 13.802 8.407 90.071 89.570 119.632 145.2- 179.5 -P6 1 3.009 14.827 8.407 89.858 89.660 118.636 144.4- 179.7 -117.9708 Pcc2 14.155 24.031 8.407 90.194 90.41 5 88.603 147.1- 179.3 -117.9708 "In all cases, cells were optimised to remove all imaginary phonon frequencies.Note that the lattice energy is not directly comparable to that calculated using the Gale and Henson potential^.'^ bEnergy per AlPO, unit. Fig.6 Comparison of the structure of the P6 (dotted line) and orthorhombic Pcc2 (solid line) unit cells obtained in this work little perturbation of the atomic coordinates and hence energy of the system. The most important result of this work is the removal of linear A1-0-P bond angles by allowing disorder of the oxygen position.The constraint of the hexagonal cell to the P6cc space group during refinement does not allow such disorder to occur, whereas the calculated lower symmetry of P6 does allow these bond angles to be optimised. Furthermore, the lower symmetry refinement in Pcc2I7 also does not impose such a constraint. Analysis of our results shows that, if the orientation of the displacement of the O(2) atom (in the P6 cell) is averaged then the average Al-O(2)-P angle gives a value close to 180": the Al-O(2)-P angles take values ranging between 145 and 155", and -145 to -155" with respect to the original O(2) position. We therefore conclude that the original speculation regarding this oxygen position is correct and that reducing the symmetry to account for this should generate a structure which gives a better fit to the experimental data.However, within this space group we cannot account for the inability to refine the neutron data with an ordered Al/P distribution." Furthermore, NMR studies'' sug-gest that the unit cell of A1P04-5 has three unique A1 and three unique P atoms. None of the suggested hexagonal models has such a unit cell. But, the orthorhombic cell has 6 T sites, suggesting that a unit cell possessing a higher (supergroup of Pcc2) symmetry may also be consistent with the diffraction data. Mora et ~1.'~ also considered other subgroups of P~cc, of which Ccc2 has a unit cell with three unique T sites. Energy minimisation of this cell results in P2 symmetry with a final energy cu.1kJ mol-' higher than the Pcc2 unit cell. It appears therefore that the Ccc2 symmetry is unstable with respect to both the Pcc2 and P6 unit cell, and does not, furthermore, result in the peak splitting noted in the X-ray diffraction pattern.17 These results show once more that there are a large number of configurations, almost equivalent in energy, on the potential-energy surface of such systems, and that with only very small atomic displacements, it should be possible to construct a unit cell which is consistent with all the experimen- tal data. Above, we have discussed how the refinement process can lead to the introduction of linear (or near-linear) A1-0-P angles. We will now consider how aspects of different potential models used in computational studies can influence the calcu- lated structures.So far, we have used the potential model of Gale and Hen~on~~ and find that this model will allow the relaxation of Al-0-P bond angles away from a linear configuration. The principle features of this model are the inclusion of a shell model representation of polarisable oxygen ion and a three-body term to represent the directional nature of the A1-0 and P-0 bonds. On the other hand, the model of van Beest et uL3' consists of only two-body terms, with no representation of oxygen polarisability. We therefore investi- gated whether this model permits the relaxation of the linear A1-0-P bond. Previous studies using this potential have shown that the parameters reproduce well the experimental structures of a number of AlPO material^.'^ We have taken our optimised hexagonal and orthorhombic cells and re-optimised them using the potential model of van Beest et uL3' In each case, the final minimum-energy structure contains the linear configuration of the Al- O(2)-P angle, this angle now having an average value of 177.9".This collinear bonding configuration will stabilise the lattice, in this case, by shielding the Coulombic repulsion between A1 and P. We note that in the calculations with the polarisable oxygen potentials29 the oxygen ions are significantly polarise$, with core-shell displacement being between 0.136 an,d 0.169 A, with the average value for the O(2) ions being 0.152 A. Furthermore, if the P6cc unit cell is optimised using the same potentials but omitting the polarisable oxygen (i.e.the oxygen is treated as a rigid ion whilst all the other parameters are kept the same), a minimum- energy configuration is found in this space group with the Al-O(2)-P angles remaining near linear. It is therefore clear that the inclusion of polarisation is important in such systems, contributing approximately 12% of the total lattice energy. The effect of the three-body term is obviously much smaller (<0.1%) but may be significant in providing a driving force J. Muter. Chem., 1996,6(11), 1837-1842 1841 to allow the minimisation to locate the non-linear Al-0-P configurations. We note similar effects when polarisation is neglected in modelling silicate materials.26 Conclusions Lattice-energy minimisation calculations using different optimisation techniques have proved to be useful in under- standing the detailed crystallographic structure of A1P04-5.We find first that reduction of the symmetry from the exper- imentally determined P6cc to P6 allows the full optimisation of the O(2) position, resulting in A1-0-P angles which are much closer to those generally observed. We calculate that the recent refinement in the orthorhombic Pcc2 space group has a slightly higher energy than the hexagonal cells, although again we note that these calculations are effectively at OK, whilst the orthorhombic cell was observed at room tempera- ture. The differences between our hexagonal unit cell and the orthorhombic unit cell of Mora et a1.17 are very small.We therefore conclude that, although the hexagonal cell is energeti- cally more favoured, entropic effects may be sufficient to stabilise the orthorhombic phase. Furthermore, the differences between the atomic positions and the magnitude of the bond distances and angles are extremely small. Thus, such a change in the space group will not affect the physical properties of the material. All these results allow us to conclude that if the position of the O(2)' [and hence the values of the Al-O(2)-P angle] could be determined with greater accuracy in the experimental refinements, a symmetry lower than P6cc is likely. The O(2) position as determined in the P6cc space group' is an average position resulting from the three-fold disorder of the atom around this point, with the actual values likely to yield more typical Al- 0-P angles (145-155'). The present calculations suggest that the hexagonal structure (P6) is closely related to the recently reported Pcc2 refinement.17 This prediction could be confirmed by following the splitting of the hexagonal peaks17 with decreasing temperature using high-resolution powder diffraction techniques. With regard to the choice of potential parameters used to describe AlPO systems, it is clear from these results that the inclusion of polarisation is of great importance in obtaining structures which do not contain anomalous Al-0-P angles.Furthermore, it is clear that careful selection of minimisation algorithm is required and that the use of lattice dynamical calculations is an invaluable aid to identifying fully stable energy minima.These calculations raise an important question pertaining to the solution of the structure of microporous materials in general: are deviations from expected T-0-T angles and optimal bond lengths a consequence of the constraints imposed by the crystallographic space group? Certainly we would consider that structure solutions containing linear T-O-T angles are likely to refine to more typical values if a lower space group is used. However, the complex nature of these materials may encourage the use of higher space groups. Nevertheless, once accurate models in high-symmetry space groups are obtained, re-refinement in lower space groups is likely to yield more characteristic geometries and reduce temperature factors introduced by the symmetry constraints. The role of lattice simulation techniques in assisting such studies is evident from this study.A. R. R-S. would like to acknowledge the support of the Cuban Embassy in London and the Royal Institution of GB for financial support during his visit to London. G.S.acknowledges MEC of Spain for financial support. The EPSRC is thanked for a ROPA grant to D. W. L. and general support. We are indebted to R. H. Jones for making the results of the work of Mora et al. available to us prior to publication. J. D. Gale is also thanked for useful discussions. References 1 S.T. Wilson, B. M. Lok, C. A. Messina, T. R. Cannan and E. M. Flanigen, J. Am. Chem. SOC.,1982,104,1146. 2 J. J. Pluth and J. V. Smith, Acta Crystallogr., Sect. C, 1986,42,283. 3 M. Ito, Y. Shimoyama, Y. Saito, Y. Tsurita and M. Otake, Acta Crystallogr., Sect. C, 1985,41,1698. 4 B. M. Lok, C. A. Messina, R. L. Patton, R. T. Gajek, T. T. Cannan and E. M. Flanigen, J. Am. Chem. SOC., 1984,106,6092. 5 P. A. Wright, R. H. Jones, S. Natarajan, R. G. Bell, J. Chen, M. B. Hursthouse and J. M. Thomas, J. Chem. SOC., Chem. Commun., 1993,633. 6 M. E. Davis, C. Saldarriga, C. Montes, J. Garces and C. Crowder, Nature (London), 1988,331,698. 7 S. I. Zones, R. A. van Norsdstand, D. S. Santilli, D. M. Wilson, L. T. Yuen and L. D. Scampani, in Zeolites: Facts, Figures and Future, ed.P. A. Jacobs and R. A. van Santen, Elsevier, Amsterdam, 1989, vol. 49, p. 299. 8 J. M. Bennett, J. P. Cohen, E. M. Flanigen, J. J. Pluth and J. V. Smith, in ACS Symp. Ser., American Chemical Society, Washington, 1983, vol. 218, p. 109. 9 S. Qiu, Q. Pang, H. Kessler and J. L. Guth, Zeolites, 1989,8,440. 10 J. W. Richardson, J. Pluth and J. V. Smith, Acta Crystallogr, Sect. C, 1987,43,1469. 11 M. P. Peeters, L. van de Ven, J. W. de Haan and J. H. C. van Hooff, J. Phys. Chem., 1995,97,9254. 12 E. de vos Burchart, H. van Bekkum, B. van de Graaf and E. T. C. Vogt, J. Chem. SOC.,Faraday Trans., 1992,88,2761. 13 A. J. M. de Man, W. P. J. H. Jacobs, J. P. W. Gilson and R. A. van Santen, Zeolites, 1992, 12, 826.14 G. Sastre, D. W. Lewis and C. R. A. Catlow, J. Phys. Chem., 1996, 100,6722. 15 N. J. Henson, A. K. Cheetham and J. D. Gale, Chem. Mater., 1996, 8,664. 16 N. Ohnishi, S. Qiu, 0. Terasaki, T. Kajitani and K. Hiraga, Microporous Mater., 1992,2, 73. 17 A. J. Mora, A. N. Fitch, M. Cole, R. Goyal, R. H. Jones, H. Jobic and S. W. Carr, J. Mater. Chem., 1996,preceding paper. 18 P. A. Vaughan, Acta Crystallogr., 1966,21,983. 19 R. E. Morris, S. J. Weigel, N. J. Henson, L. M. Bull, M. T. Janicke, B. F. Chmelka and A. K. Cheetham, J. Am. Chem. SOC., 1994, 116,11849. 20 Computer Simulation of Solids, ed. C. R. A. Catlow and W. C. Mackrodt, Springer-Verlag, Berlin, 1982,vol. 166. 21 J. D. Gale, GULP program, Royal Institution of GB and Imperial College, London, 1991-1996. 22 P. P. Ewald, Ann. Physik, 1921,64,253. 23 D. F. Shanno, Math. Comp., 1970,24,647. 24 J. Simons, P. Jsrgensen, H. Taylor and J. Ozment, J. Phys. Chem., 1983,87,2745. 25 R. A. Jackson and C. R. A. Catlow, Mol. Sim., 1988,1,207. 26 R. G. Bell, R. A. Jackson and C. R. A. Catlow, J. Chem. SOC.,Chem. Commun., 1990,782. 27 P. A. Wright, S. Natarajan, J. M. Thomas, R. G. Bell, P. L. Gai- Boyes, R. H. Jones and J. S. Chen, Angew. Chem., Int. Ed. Engl., 1992,31,1472. 28 M. D. Shannon, J. L. Casci, P. A. Cox and S. J. Andrews, Nature (London),1991,353,417. 29 J. D. Gale and N. J. Henson, J. Chem. SOC.,Faraday Trans., 1994, 90, 3175. 30 B. G. Dick and A. W. Overhauser, Phys. Rev., 1958,112,90. 31 B. W. H. van Beest, G. J. Kramer and R. A. van Santen, Phys. Rev. Lett., 1990,64, 1955. 32 Modelling of Structure and Reactivity in Zeolites, ed. C. R. A. Catlow, Academic Press, London, 1992. Paper 6/04149C; Received 13th June, 1996 1842 J. Mater. Chem., 1996, 6(11), 1837-1842

ISSN:0959-9428    

DOI:10.1039/JM9960601837

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Structure of Zn( O3PC2H4CO2H)*0.5CsH5NH2 and XANES-EXAFS study of the intercalation of amines into Zn (03PR)=H20zinc alkylphosphonates StCphanie Drumel," Pascal Janvier,b Martine Bujoli-Doeuff" and Bruno Bujoli *' "IMN, UMR CNRS 110, Facultk des Sciences et des Techniques, 2, rue de la Houssinisre, 44072 Nantes Cedex 03, France bLaboratoire de Synthsse Organique, URA CNRS 475, Facultk des Sciences et des Techniques, 2, rue de la Houssinisre, 44072 Nantes Cedex 03, France On the basis of XANES-EXAFS experiments, neither the dehydration nor the subsequent n-alkylamine intercalation in Zn(O,PCH,)-H,O appears to be topotactic. On the contrary, the whole process consists of breaking Zn-0 bonds present in the hydrated material, so that no bridging oxygen remains in the inorganic sheet.This hypothesis is supported by the structuralo determination of an aniline intercalate: Zn(0,PC2H4C02H).0.5C6H5NH2[orthorhombic, space group Pbcn, a =29.880( 6) A, b=8.526(2) A,c= 14.720(3)A,V=3750.3(8) A3, Z= 16, R=0.043 and R,=0.047; 2063 observed reflections, I >20(1)]. For steric reasons, only half of the zinc atoms are coordinated to aniline; the second half of the metal atoms that are not bound to the amine retain the environment present in the initial anhydrous phase. The chemistry of phosphonates has received increasing atten- tion over the last five years, with the production of porous materials,' self-assembled monolayers or multilayered mate- rials,2 with properties such as magnetic materials,, non-linear optical material^,^ ion exchangers, sorbents5 and catalysts.6 Parallel to this, many studies were reported relating the intercalation properties of layered divalent metal alkyl- and aryl-phosphonates of composition M"(O,PR)-H,O. Apart from the case of copper," all these materials (M =Mn, Co, Zn, Cd, Mg, Fe, Ni) have the same in-plane structure,8 with an octahedral environment of oxygen atoms around the metal atom, in which one of the vertices is occupied by the water molecule.Cao and Mallouk,' Clearfield et al." and Cao et have demonstrated that this water molecule can be removed thermally and that, in certain cases (M =Co, Zn, Cd), it is possible to intercalate alkylamines in the resulting anhydrous phase to yield M"(O,PR).R"H;,.In all these studies, the process was thought to be topotactic, and the nitrogen atom of the alkylamine was assumed to occupy the vacant coordi- nation site created after dehydration of the phosphonate hydrate, the structure of the layer being roughly retained. However, in a recent publication by Poojary and Clearfield," the structures of three members of this family of intercalates, Zn1r(03PC6H5).R'NH2(R'=C3H7, C4H9, C5H11) were solved ab initio from X-ray powder diffraction data and refined by Rietveld methods. This study showed that the zinc atoms were four-fold coordinated by three phosphonate oxygens and one nitrogen atom, giving evidence of a strong structural rearrange- ment within the layer during the amine intercalation step.However, the authors concluded that this rearrangement was possibly limited to the case of the bulky phenylphosphonate groups, which do not allow enough space for the alkylamine to intercalate. They also mentioned that a structural determi- nation would be helpful to learn if the intercalation process was really topotactic in the zinc n-alkylphosphonates [i.e. Zn(O,PCH,)], as assumed by Cao and Mallouk.' Moreover, no hypothesis that the dehydration step could itself be non- topotactic was considered. In an independent work, we have noticed that the IR spectra [in the v(NH) and v(PO,) regions] of Zn(03PC2H4NH2)ld and Zn(O3PCH3).n-C4H9NH2 were nearly superimposable, suggesting that the two compounds had similar arrangements within the layers, and consequently that the amine intercalation in Zn(O,PCH,) was not topotactic either, as in the phenyl analogue.This hypothesis was con- firmed by EXAFS and XANES studies, which are reported in this paper. Experimenta1 Synthesis and characterization Zn(03PCH,).H20, Zn(O,PCH,), Zn(O,PCH,).n-C,H,NH, and Zn(03PC2H4NH2) were prepared as described pre- vio~sly.'~~'Single crystals of Zn(0,PC2H,C02H).0.5C6H5NH2 were synthesized by placing a mixture of zinc nitrate (1 mmol), 2-carboxyethylphosphonicacid ( 1mmol) and aniline (6 mmol) with 16 ml water in the Teflon cell (80% full volume) of an autoclave, which was then sealed and left at 110°C in a dry- ing oven for 6 days. Zn"(O3PC2H4CO2H)~O.5H2NC6H5was obtained as white crystals, in 79% yield.The chemical analyses, IR and 31PMAS NMR data were identical for the compounds, prepared by direct synthesis or by reaction of aniline with Zn"(03PC2H4C02H).H2012or Zn" (03PC2H4C02H). The IR absorption spectra (4000-400 cm-') were obtained by using a FTIR Nicolet 20SX spectrometer with the usual KBr pellet technique. Extended X-ray absorption fine structure (EXAFS) and K-edge spectra were recorded at L.U.R.E (the French National Synchrotron Radiation facility) on the EXAFS-I11 station using the X-ray emission of a positron beam (1.85 GeV; 150 mA) in the storage ring DCI. The samples were prepared as cellulose pellets and studied on the EXAFS- I11 spectrometer, equipped with a two-crystal Si( 31 1) mono-chromator, using air-filled ionization chambers.The X-ray absorption near-edge structure (XANES) (cf. EXAFS) spectra were recorded in the transmission mode, step by step every 0.3 eV (2 eV) with 1s (2 s) accumulation time per point. The XANES spectrum of a zinc metallic foil was recorded after or just before an unknown XANES spectrum of a sample to check energy calibration. The background absorption was calculated by using a theoretical expression developed by Lengeler and Eisenberger.', The single atomic absorption of the absorber was interpolated by a fifth-degree polynomial. The radial distribution function was obtained by a FouFier transform of the k3.~(k)spectra (k is the wavevector in A-' and x is the EXAFS contribution of the absorption spectrum using a Kaiser window contribution; z =3.5).The basic analysis J. Muter. Chem., 1996,6(11), 1843-1847 1843 Table 1 Selected bond lengths/A and angles/degrees for the non-hydrogen atoms of Zn'1(03PC,H4C0,H)~0.5C,H,NH2 Zn ( 1)-0( 12)" 1.936( 6) P( 1)-O( 11) 1.51 l(6) Zn( 1 )-O( 21)" 1.928(5) P( 1)-O( 12) 1.493(6) Zn( 1)-O( 13)b 1.995(5) P( 1)-O( 13) 1.567(5) Zn( 1)-Nu 2.023 (7) P(1)-C( 11) 1.808( 8) Zn( 2)-O( 11 )" 1.948( 6) P(2)-O( 21 )" 1.529(6) Zn( 2)-O( 22)" 1.945(5) P(2)-O( 22) 1.5 19 (6) Zn( 2)-O( 1 3)b 1.989(5) P(2)-O( 23) 1.514(6) Zn( 2)-O( 23)b 1.901 (6) P(2)-C( 21 )" 1.808 (9) O(12)"-Zn( 1)-O(21)" 113.6(2) O(12)"-Zn( 1)-O( 13)b 102.6(2) O(12)"--Zn( 1)-N" 103.5( 3) 0(2l)"-Zn(l)-O( 13)b 103.6(2) 0(21 T-Zn (1)-Nu 112.5 (3) O(13)b-Zn( 1)-Nu 121.0(3) O(11)"-Zn(2)-0(22)" 106.8(2) 0(11)"-Zn(2)-0(13)b 109.3(2) O(11)"-Zn(2)-0(23)b 108.8( 2) 0(22)"-Zn(2)-0( 13)b 103.2(2) 0(22)1-Zn(2)-0(23)b 114.4(2) O(13)b-Zn( 2)-O( 23)b 114.1 (2) "Atom related by -1/2+x, 1/2-y, 1-z; bl--~,l-y, 1-2; "x, l-y, -1/2+z.of the EXAFS spectra followed standard numerical procedures already detailed e1~ewhere.l~ The first shell was extracted by a back Fourier transform including a removal of the Kaiser window contribution, then studied through a fitting procedure with simplex and gradient methods using the MINUIT pro- gram and MacKale tabulated amplitudes and phase shifts. The reliability of the fit was determined by the error factor p= C [kxe,,(k) -kxthe0(k)l2/C.[kxeXp(k)l2. For the five compounds reported, the fitting procedure allowed first CN, then r, and the Debye-Waller factor to vary iteratively.31P Solid-state MAS NMR spectra were recorded at room temperature on a MSL-400 NMR Bruker spectrometer at 161.98 MHz. Samples were spun in an Andrew-type rotor at ca. 8 kHz. Chemical shifts were externally referenced to 85 mass% H3PO4. Structure determination of Zn(0,PC2H,C02H)*0.5C,H2NH2 A colourless platelet of Zn(03PC2H4C0,H).0.5C,H,NH2NH2of approximate dimensions 0.02 x 0.20 x 0.28 mm3 was mounted on a glass fibre. All measurements were made on a Siemens P4 diffractometer wjth graphite-monochromated Mo-Ka radi- ation (1= 0.710 73 A). Cell constants and an orientation matrix for data collection were obtained from least-squares refinement, with use of the setting angles of 25 randomly oriented reflec- tions in the range 10<28/degrees <as, correspondtng to an orthorkombic cell [a 29.880( 6) A, b =8.526(2) A, c = 14.720(3) A, V=3750.3(8) A3, Z= 16, pcalc=1.874 g ~m-~, p= 27.8 cm-'1. To check on crystal and instrument stability, three representative reflections were measured every 60 min and no decay was observed.An empirical absorption correction based on $-scan measurements was applied and the data were corrected for Lorentz and polarization effects. The data were collected to 60" in 28 using the 0-28 scan technique. On the basis of the systematic absences and the successful refinement of the structure, the space group was found to be Pbcn.The atomic scattering factors were taken from Cromer and Waber,l5 and anomalous dispersion corrections were taken from Cromer The positions of the zinc and phosphorus atoms were determined from a three-dimensional Patterson map, with the oxygen, nitrogen and carbon atoms being found from suc- cessive difference Fourier maps. The non-hydrogen atoms were refined anisotropically. The final cycle of full-matrix least- squares refinement for 244 variables converged (largest parameter shift was 0.03 times its e.s.d.) with unweighted and weighted agreement factors of R =CIIF, I -IF,ll/ClF, I =0.043 and R,= [Zw(lF,I -~Fc~)2/C~(F,2)]1/2=0.047,where w= 4Fo2/[o(Fo2)]" Selected bond lengths and angles are given in Table 1. Atomic coordinates, thermal parameters, and bond lengths and angles have been deposited at the Cambridge Crystallo- graphic Data Centre (CCDC).See Information for Authors, J. Muter. Chern., 1996, Issue 1. Any request to the CCDC for this material should quote the full literature citation and the reference number 1145/10. Results and Discussion In the structure of the zinc methylphosphonate Zn(O,PCH,).H,O, the zinc atoms have a distorted octahedral coordination. Five oxygens of the phosphonate groups are connected to the metal atom, while the sixth coordination site is occupied by a water molecule. It was assumed by previous authors that the anhydrous zinc methylphosphonate was iso- morphous with the hydrated phase and that no changes in the layer arrangement occurred, except a reduction of the coordi- nation number to five.Upon exposure to linear primary amines, intercalation occurred and the amine was supposed to occupy the same coordination site as the water molecule in the monohydrate. However, some doubts about this hypothesis appeared when we compared the IR spectra of the n-butyl intercalate Zn(O3PCH3).n-C4HgNH, [v(NH) 3293, 3265 and 3177 (m); v(P03) 1120 (s), 1057 and 1040 (vs), 1008 (s)] with that of our previously reported Zn(03PC2H4NH2) [v(NH) 3291, 3265 and 3180 (m); v(P03)1119 (s), 1060 and 1040 (vs), 1005 (s)],'~ the X-ray structure of which shows that the zinc atoms are tetrahedrally coordinated with three oxygens of phosphonate groups and the nitrogen atom of the aminoethyl chain bound to phosphorus.In fact, as the two spectra are nearly superimposable, it is probable that the arrangements within the layer are similar for the two compounds, and the IR spectrum in the v(NH) region gives clear evidence of the coordination of the nitrogen atom of the n-butylamine to the zinc atoms, in Zn(0,PCH3).n-C4HgNH2. XANES-EXAFS measurements [ZnO (reference), Zn(O,PCH,).H,O, Zn(O,PCH,), Zn(03PCH3).n-C4H,NH, and Zn(O,PC,H,NH,)] were then carried out to confirm this point. In the K-edge spectra of zinc@) compounds, no prepeak is observed, since no transition is allowed (d10 configuration with fully occupied orbitals) before the absorption edge, which corresponds to the first allowed transition 1s to Tlu. If we compare the absorption edge of Zn(03PCH3)-H20 (six-coordi- nate metal), with a maxima at 9668 eV, with that of ZnO or Zn(03PC2H4NH,), a significative decrease of the intensity is and Ibers.16 For the data reduction, structure solution and observed for the two latter compounds, consistent with a refinement, the SHELXTL PLUS package was used on the tetrahedral environment for the metal atoms, and we can note basis of 2063 independent reflections corresponding to the same behaviour for both Zn(03PCH,).n-C4H9NH, and the condition I> 241).Zn(03PCH3) (Fig. 1). Table 2 Results of the fits for the first zinc coordination shell in Zn(O,PC,H,NH,), compound ~~ atom (CN)" r/Ab (a& atom (CN)" Zn(O3PC2H4NH2) Zn(O3PCH3)* n-C4HgNH2 Zn(03PCH,) -N (1) N (1) 2.04 (0.06) -1.97 (0.07) 0 (3) 0 (4) 0 (3) Zn(0,PCH,)wC4H9NH2 and Zn(03PCH3) r/Ab (a/A)" AE,/eVd A"/.)" 1.93 (0.06) 6.4 0.49 1.97 (0.07) 7.0 0.37 1.95 (0.07) 6.8 0.41 "Coordination number.bZn atom distance. "Debye-Waller factor. The relatively high value of the o factor is related to the dispersion of the Zn-0 distances in the Zn03N and Zn04 tetrahedra, for which an average Zn-0 distance is given. dEdge variation. 'Reliability factor of the fit. 1844 J. Muter. Chern., 1996, 6(11), 1843-1847 value close to four, for which the refinement was satisfactory I(Fig. 2), in agreement with the Zn K-edge absorption spectrum. 1.52-r1 N.--9600 9650 9700 9750 9800nrL I 0.5 j. 9600 9700 9750 9800 energylev Fig.1 Comparison of the X-ray absorption zinc K-edge spectra (a) of Zn(O,PCH,).H,O (-) and Zn(0,PC2H4NH,) (0);(b)of Zn(O,PCH,) (-), Zn(O,PCH,).n-C,H,NH, (A)and Zn(O,PC,H,NH,) (0) If we now examine the Fourier transform obtained for the four zinc phosphonates, the first well resolved peak corresponcs to the first coordination sphere of the metal atom (Rz2A). The final simulations of the first EXAFS peaks led to the data summarized in Table 2. The metal-ligand distances in Zn(03PC2H4NH2) were consistent with the corresponding crystallographic values and were then used to refine successfully the spectrum of Zn(0,PCH,)-n-C,H9NH2. As expected, a six- fold coordination of oxygens was found for Zn(O,PCH,)-H,O. In the case of the anhydrous form Zn(O,PCH,), all attempts to fit the EXAFS spectrum with a coordination number (CN) of six or five invariably resulted in a drop in the CN to a 2 4 6 8 10 12 RIA Fig.2 Modulus of the Fourier transform of the k~(k)data at the Zn K-edge EXAFS spectrum for Zn(O,PCH,) (uncorrected for phase shift) Two important conclusions may be drawn from these XANES-EXAFS measurements.(i) The n-butylamine intercal- ation in Zn(03PCH3).H20 does not occur by the replacement of the coordinated water molecule by the nitrogen of the amine, because the zinc atoms are four-coordinate in the intercalate. Moreover, the strong similarities observed in the IR, XANES-EXAFS and 31P MAS NMR data for Zn(O,PC,H,NH,) (,'P ais0 26.7) and Zn(O,PCH,)-n- C,H,NH, (,'P ais0 27.1) leaves no doubt about the presence of common inorganic frameworks in the layers of the two products.Additional evidence for this is given by the recent report by Clearfield et ul.ll of the structures of zinc phenylphos- phonate amine intercalates, in which the arrangement within the layers are identical to that of Zn(O,PC,H,NH,) (Fig. 3). The only difference between Zn(03PR).n-C4H9NH2 (R =C,H,, CH,) and Zn(0,PC2H,NH2) is that the nitrogen atom of the Fig. 3 Comparison of the layer arrangement in (a) Zn(O,PC,H,).n-C,H,NH, and (b)Zn(0,PC,H,NH2). The carbon atoms have been omitted for clarity. J. Muter. Chem., 1996, 6(11), 1843-1847 1845 alkylamine is bound to zinc in the former compounds, and comes from the amino end of the alkyl chain borne by phosphorus of the upper and lower layers for the latter compound, leading to a three-dimensional structure. (ii) A coordination number of four is observed for the anhydrous phase Zn(O,PCH,), instead of the five-fold coordination pos-tulated in previous studies.This means that a rearrangement occurs even in the dehydration step, which is consequently not topotactic. In the copper alkyl- and phenyl-phosphonates Cu(O,PR).H,O, we have demonstrated that the water loss resulted in a structural rearrangement17 because of the highly unfavourable coordination environment created by dehy- dration. In the case of Zn(O,PCH,)-H,O, the loss of coordi- nated water would lead to a distorted square-pyramidal site for the zinc atom, which has never been observed until now in the chemistry of zinc phosphonates.Therefore a structural rearrangement was expected, and we wished to know if a decrease of the coordination number was associated with this phenomenon. If we compare the framework of the layers in Zn(O,PC,H,NH,) or Zn(O,PC,H,)-n-RNH, [Fig. 4(C), in which the nitrogen atom, coordinated to the metal atom has been artificially removed] and Zn(O,PCH,).H,O [Fig. 4(A), in which the water molecule has been artificially removed], we can see that the linkage in the former compound is obtained simply by breaking Zn-0 bonds as indicated in Fig. 4(B), so that no bridging oxygen remains in the inorganic sheet. Then, by expanding the structure along the 2 axis of the layer, we (A) b Fig.5 Schematic representation of a Zn"(O,PC,H,CO,H)-O.SC,H, NH, layer as seen perpendicular to the a axis. The carbon atoms have been omitted for clarity. O(13) 1)O(21) Fig. 6 Schematic representation of the coordination about the two Fig. 4 Schematic representation of the rearrangement occuring during types of zinc atoms in Zn"(O,PC,H,CO,H)~O.SC,H,NH,,and the the dehydration/amine intercalation process in Zn(O,PCH,).H,O numbering scheme used in the Tables 1 and 2 1846 J. Muter. Chem., 1996, 6(11), 1843-1847 a C Fig. 7 Schematic representation of the layer arrangement in Zn"(O,PC,H,CO,H)~O.SC,H,NH,,as seen perpendicular to the b axis can see that the two arrangements are identical, with large 16-membered rings (noted AlB2C3D4), connected by smaller 8-membered rings (noted B2D4).Consequently, the four-fold coordination of Zn(O,PCH,) would simply result from the removal of the water molecule together with a tilt of the phosphonate groups, in such a way that only one of the phosphonate oxygens bridges zinc atoms. This modification in the intralayer linkage would not induce significant differ- ences in the unit-cell parameters of the corresponding product. In support of this hypothesis, the behaviour of Zn(0,PC2H4C02H)-H20 towards amine intercalation is very interesting. We have demonstrated earlier that Zn(03PC2H4C02H).H2012was a structural analogue of the n-alkyl Zn(03PR)-H20 series. When this compound or its dehydrated form, Zn(03PC2H4C02H), was suspended in acetonitrile with aniline, Zn(0,PC2H4C02H).0.5C6H2NH, was obtained.This structure is layered (Fig. 5), with two different tetrahedral environments for the zinc atoms (Fig. 6): the first one consists of four phosphonate oxygens, the second is made of three phosphonate oxygens and the nitrogen atom of aniline. There are two types of RPO, groups, one of them being bonded to three zinc atoms through each of its oxygen atoms. The other kind of phosphonate units are connected to four zinc atoms, with one of the oxygens bridging two metal atoms. The phenyl rings and the 2-carboxyethyl chains are extended roughly perpendicular to the layers (Fig. 7). As the C02H group of the alkyl chain remains unchanged after intercalation of aniline, it was reasonable to assume that the dehydration/intercalation steps were identical to those described for Zn(03PCH3).H20 with n-alkylamines, except that only half of the zinc atoms in Zn(0,PC2-H4C02H).0.5C6H2NH2were coordinated by aniline, owing to the bulky phenyl group of the amine which did not allow a full occupation of the available sites.The second half of the zinc atoms, which are not coordinated to the amine, would then retain the environment present in the initial anhydrous phase (ZnO, tetrahedra), as well as half of the RPO, groups in which one bridging oxygen is present. References 1 (a)G. Alberti, F. Marmottini, S. Murcia-Mascaros and R. Vivani, Angew. Chem., Int. Ed. Engl., 1994, 33, 1594; (b) G. Alberti, U.Costantino, F. Marmottini, R. Vivani and P. Zapelli, Angew. Chem., Int. Ed. Engl., 1993, 32, 1557; (c) J. Le Bideau, C. Payen, P. Palvadeau and B. Bujoli, Inorg. Chem., 1994, 33, 4855; (d) S. Drumel, P. Janvier, D. Deniaud and B. Bujoli, J. Chem. SOC., Chem. Commun., 1995, 1051; (e) K. Maeda, J. Akimoto, Y. Kiyozumi and F. Mizukami, Angew. Chem., Int. Ed. Engl., 1995, 34, 1199; (f) K. Maeda, J. Akimoto, Y. Kiyozumi and F. Mizukami, J. Chem. SOC., Chem. Commun., 1995,1033. 2 G. Cao, H. Hong and T. E. Mallouk, Acc. Chem. Res., 1992, 25, 420. 3 B. Bujoli, 0. Pena, P. Palvadeau, J. Le Bideau, C. Payen and J. Rouxel, Chem. Muter., 1993, 5, 583; S. G. Carling, P. Day, D. Visser and R. K. Kremer, J. Solid State Chem., 1993, 106, 111; V.Soghomonian, R. Diaz, R. C. Haushalter, C. J. OConnor and J. Zubieta, Inorg. Chem., 1995,34,4855. 4 S. B. Ungashe, W. L. Wilson, H. E. Katz, G. R. Scheller and T. M. Putvinski, J. Am. Chem. SOC.,1992,114,8717. 5 A. Clearfield, New Developments in Ion Exchange Materials, Kodansha, Ltd., Tokyo, 1991. 6 D. Deniaud, B. Schollhorn, D. Mansuy, J. Rouxel, P. Battioni and B. Bujoli, Chem. Muter., 1995,7, 995. 7 Y. Zhang and A. Clearfield, Inorg. Chem., 1992,31,2821. 8 (a)G. Cao, H. Lee, V. M. Lynch and T. E. Mallouk, Inorg. Chem., 1988, 27, 2781; (b)G. Cao, V. M. Lynch and T. E. Mallouk, Solid State Ionics, 1988, 26, 63; (c) K. J. Martin, P. J. Squattrito and A. Clearfield, Inorg. Chim. Acta, 1989, 155, 7; (d) G. Cao, V. M. Lynch and L. N. Yacullo, Chem. Muter., 1993,5, 1000. 9 G. Cao and T. E. Mallouk, Inorg. Chem., 1991,30,1434. 10 Y. Zhang, K. J. Scott and A. Clearfield, J. Muter. Chem., 1995, 5, 315. 11 D. M. Poojary and A. Clearfield, J. Am. Chem. SOC., 1995, 117, 11278. 12 S. Drumel, P. Janvier, P. Barboux, M. Bujoli-Doeuff and B. Bujoli, Inorg. Chem., 1995,34, 148. 13 B. Lengeler and P. Eisenberger, Phys. Rev. B, 1980,21,4507. 14 A. Michalowicz, Logiciels pour la Chimie, SociCtC Frangaise de Chimie, Paris, 1991, p. 102. 15 D. T. Cromer and J. T. Waber, International Tables for X-ray Crystallography, Kynoch Press, Birmingham, 1974, vol. IV, Table 2.2B. 16 D. T. Cromer and J. A. Ibers, International Tables for X-ray Crystallography, Kynoch Press, Birmingham, 1974, vol. IV, Table 2.3.1. 17 J. Le Bideau, B. Bujoli, A. Jouanneaux, C. Payen, P. Palvadeau and J. Rouxel, Inorg. Chem., 1993,32,4617. Paper 6/02419J; Received 9th April, 1996 J. Muter. Chem., 1996,6(11), 1843-1847 1847

ISSN:0959-9428    

DOI:10.1039/JM9960601843

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Intercalation of 2-, 4=sulfanyIpyridine,2,2'-and 4,4'-dithiobispyridine into VOP04 and gel-V,O, interlayer spaces Teruyuki Yatabe and Gen-etsu Matsubayashi* Department of Applied Chemistry, Faculty of Engineering, Osaka University, Machikaneyama 1-1 6, Toyonaka, Osaka 560, Japan 2- (2-Spy) and 4-sulfanylpyridine (4-Spy) have been inserted into the gel-V205 interlayer space, accompanied by oxidative S-S coupling, to form intercalation compounds including mostly N-protonated 2,2'- (2,Y-pySSpy) and 4,4'-dithiobipyridine (4,4'-pySSpy) moieties as a guest. 2,2'-pySSpy and 4,4'-pySSpy were also intercalated into the gel-V205 interlayer space to yield similar intercalation compounds. 2-Spy, 4-Spy and 4,4'-pySSpy formed no stable intercalation compounds with VOP04 solids, while 2,2'-pySSpy was inserted into the VOP04 lattice to afford an intercalation compound with fewer N-protonation sites.X-Ray diffraction patterns and X-ray photoelectron spectra of the intercalation compounds are discussed. Introduction Interlayer spaces of lamellar inorganic solids afford unique fields for reactions, which are different from those of liquid Lamellar oxovanadium(v) phosphate (VOP04) and vanadium(v) oxide xerogel (gel-V205) compounds can include various organic3-' and organometallic cationic compounds in the interlayer spaces, accompanied by some reduction of the host Pyridine and its derivatives are intercalated into the layered spaces and are protonated at nitrogen.17*18 Since these vanadium(v) compounds behave as oxidizing reagents, they can be intercalated with aniline, pyrrole and their derivatives through oxidative polyrnerizati~n.~'-~~ Thiol compounds are subject to oxidative S-S coupling to afford disulfide compounds.4-Sulfanylaniline can be oxidized by VOP04 2H20 to give the 4,4'-dithiobis(4-anilinium) cation in the interlayer space.25 This study reports the intercalation of 2- and 4-sulfanylpyrid- ine into the VOP04 and gel-V205 interlayer spaces as well as that of 2,2'- and 4,4'-dithiobipyridine. The arrangements and electronic states of the guest molecules in the interlayer spaces as well as the N-protonation of the guest molecules are discussed on the basis of X-ray diffraction patterns and X-ray photoelectron spectra. Experimental Materials The layered compounds VOP04 QH2017 and gel-V205* l.6H2026 were prepared according to the literature procedures.2- (2-Spy), 4-sulfanylpyridine (4-Spy), 2,2'- (2,2'- pySSpy) and 4,4'-dithiobipyridine (4,4'-py SSpy) were commer- cially available. Intercalation of 2-and Csulfanylpyridine into the V,05 interlayer space Finely powdered gel-V205 -1.6H20 (300 mg, 1.4 mmol) was added to an ethanolic (40cm3) solution of 2-Spy (510mg, 4.5 mmol) and the suspended solution was stirred at room temperature for 5 days in the dark. The resulting solids were collected by centrifugation, washed with acetone several times Found: C, 11.4; H, 1.5; N, 2.4%. Calc. for C2.5H3.g No.506.oSo.5V2: C, 11.8; H, 1.55; N, 2.75%. The water contents of 1 and 2 were determined by thermo- gravimetry.The presence of 2,2'-pySSpy and 4p'-pySSpy moieties in the V205 interlayer space was confirmed as follows. These intercalation compounds were dissolved in NaOH aque- ous solutions, which were extracted by dichloromethane, fol- lowed by removal of the solvent to afford organic solids. They exhibited parent peaks at m/z 221 in the FAB mass spectra due to 2,Y-py SSpy and 4,4'-pySSpy. These disulfide compounds were confirmed by the 'H NMR spectra of their C2Hl] chloroform solutions. Intercalation of 2,2'-and 4,4'-dithiobipyridine into the V205 interlayer space Finely powdered gel-V205 l.6H20 (300 mg, 1.4 mmol) was added to an ethanolic (25 cm3) solution of 2,Y-pySSpy (940 mg, 4.3 mmol) and the suspended solution was stirred for 5 days at room temperature in the dark.The resulting solids were collected by centrifugation, washed with acetone and dried in vucuo to give V205( H20)0.25( 2-C5H4NSS-2'-C5H4N)o.2 3. Found: C, 10.0; H, 1.1; N, 2.4%. Calc. for C2.0H2.1 NO.405.3S0.4V2: C, 10.4; H, 0.9; N, 2.4%. Similarly, an intercalation reaction of powdered gel- V205* l.6H20 with 4,4'-pySSpy in ethanol gave V205(H20)o~7(4-C5H4NSS-4'-C5H4N)o~24. Found: C, 9.7; H, 1.3; N, 2.1%. Calc. for C2~oH,~oNo~405~7So~4V2: C, 10.0; H, 1.3; N, 2.3%. Intercalation of 2-and 4-sulfanylpyridine into the VOP04 interlayer space In an ethanolic (40 cm3) solution of 2-Spy (500mg, 4.5 mmol) finely powdered VOP04 -2H20 (300 mg, 1.5 mmol) was sus-pended and the solution was stirred at room temperature for 3 days.The resulting solids were collected by centrifugation and washed with acetone several times. They exhibited no layered structure based on their X-ray diffraction spectra. A similar intercalation reaction of 4-Spy did not result in the production of a layered compound. C5H4N)0.21. Found: C, 10.6; H, 1.3; N, 2.35%. Calc. for C2~oH3,1No.405~7So.4V2:C, 10.05; H, 1.3; N, 2.35%. Similarly, 4-Spy was reacted with gel-V205 l.6H20 to afford an intercal- ation compound V205( H20)0.g5( 2.4-C5H4NSS-4'-C5H4N)o.25 Intercalation of 2,2-and 4,4'-dithiobipyridine into the VOP04and dried in uucuo to afford V205(H20)0.75(2-C5H4NSS-2'-interlayer space Finely powdered VOPO, -2H20 (300 mg, 1.5 mmol) was dis- persed in an ethanol (40 cm3) solution of 2,2'-pySSpy (1.0 g, 4.5 mmol) and the solution was stirred at room ternper-J.Muter. Chern., 1996,6(11), 1849-1852 1849 ature for 5 days. The solids obtained was collected by centrifugation, with acetone and dried in vacuo to give VOPO4(H20)0,3(EtOH),.,( 2-C5H4NSS-2'-C5H4N)o.15 5. Found: C, 16.9; H, 2.7; N, 1.8%. Calc. for C3.3H7.2No.306.2PSo.3VC, 16.4; H, 3.0; N, 1.7%. The solids showed two kinds of 001 reflection peaks in the X-ray diffrac-tion patterns, as described later; one was due to the VOPO, lattice intercalated with 2,2'-pySSpy moieties and the other was due to the VOPO, interlayer having water and ethanol molecules. The latter peak was confirmed by the peak of the solids obtained by stirring an ethanolic solution of powdered VOPO, 2H20 for 5 days.Reaction of 4,4'-pySSpy with dispersed VOP04*2H20in ethanol under similar conditions afforded only the VOP04 solids containing water and ethanol molecules without inter-calation of 4,4'-pySSpy. Physical measurements IR, powder X-ray diffraction patterns and X-ray photoelectron spectra (XPS) were measured as described previo~sly.~~'H NMR spectra were recorded at 270 MHz using a JEOL JNM-EX 270 spectrometer at the Faculty of Science, Osaka University, and the chemical shifts were measured relative to tetramethylsilane as an internal standard in C2Hl]chloroform. FAB mass spectra were measured using a JEOL DX 303-HF spectrometer. MO calculations 2,2'-Py SSpy and 4,4'-pySSpy and their N-protonated species were geometrically optimized by the PM3 method27 using the MOPAC version 6.03 program2* on an NEC SX-3R supercomputer.Results and Discussion Intercalation of sulfanylpyridine and dithiobipyridine compounds into the V205and VOPO, interlayer spaces 2-Spy and 4-Spy are inserted into the gel-V205interlayer space to give intercalation compounds 1 and 2. Fig. 1 shows their X-ray diffraction patterns together with that of V205-l.6H20. These intercalation compounds have layered structures, the interlFyer distances of 1 and 2 being estimated to be 12.4 and 13.2 A, respectively. Hence, their interlayer spaces are enlarged n0°' 003 JLL002 J\+ by 3.6 and 4.4 A,respectively, compared with that (8.8 A)of V205.0.6H20.29 The guest molecules consist of mostly N-protonated 2,2'-pySSpy and 4,4'-pySSpy moieties.The IR spectra of these intercalation compounds exhibited broad v(N-H) and v(C-H) stretching bands in the 3000-3300 cm-' region characteristic of the N-protonated pyridyl moiety and bands at 1280 and 1330 cm-', due to the 2,2'-pySSpy and 4,4'-pySSpy skeletons, respectively. No v( S-H) bands were visible in the 2400-2600 cm-' region. The S-S coupled guest mol-ecules have been confirmed by FAB-MS and 'H NMR spectra of the organic compounds extracted from the V,05 host lattice, as described in the Experimental section. Protonation of the pyridyl nitrogen atoms is deduced from XPS of the intercal-ation compounds, as described below.These findings indicate oxidative S-S coupling of the guest moieties caused by the V205 host as an oxidant, followed by the protonation of pyridyl nitrogen atoms (Scheme 1). On the other hand, reac-tions of VOPO, *2H20with 2-Spy and 4-Spy afforded no intercalation compounds with layered structures. 2,2'-Dithiobipyridine can be inserted into the V205 and VOP04 interlayer spaces to give intercalation compounds 3 and 5. Although 3 exhibited the X-ray diffraction patterns due to the 2,2'-pySSpy-V205 intercalation compound, 5 gave peaks due to the VOPO, lattice intercalated with the 2,2'-pySSpy moieties, together with those due to intercalated water and ethanol molecules. Fig. 2 shows the X-ray diffraction patterns of the 2,2'-pySSpy-VOPO, intercalation compounds, together with that of the VOPO, solids containing water and ethanol molecules.The solids obtained after the reaction of VOPO, *2H20with 2,2'-pySSpy in ethanol for 5 days afforded two 001 reflection peaks (28= 5.5 and 9.6") due to the VOP04 Scheme 1 I I * 1 I 5.0 10.0 15.0 20.0 25.0 28ldegrees I ~ ~~~ 10.010.0 15.015.0 20.020.0 2525.O.O Fig. 2 X-Ray diffraction patterns of the 2,2'-pySSpy-VOPO, intercal-28idegrees ation compound obtained (a) after reaction for 5 d, (b) after reaction for 10 d, and (c) the solids obtained by stirring of an ethanolic solution suspended with powdered VOP04.2H,0 for 5 d. The intercalationFig. 1 X-Ray diffraction paterns of (a) V205(H20)o.75(2-CSH4NSS-2 and compounds exhibit reflection peaks corresponding to the interlayer2'-C5H4N),,, 1, (b) V205(H20)0.95(4-C5H4NSS-4/-C5H4N)o.25 (c) V205* l.6H20 spaces containing 2,Y-pySSpy and ethanol/water without 2,2'-pySSpy. 1850 J.Muter. Chern., 1996, 6(11), 1849-1852 lattice intercalated with the 2,2’-pySSpy moiety and the lattice containing water and ethanol molecules, respectively [Fig. 2(a)]. The latter peak corresponds to the VOPO, solids containing no 2,2’-pySSpy moieties [Fig. 2(c)]. Even after reaction for 10 days the latter peak was still visible [Fig. 2(b)]. The obtained intercalation compound has layers containing both 2,2’-pySSpy moieties and ethanol/water molecules. The interlayer spacings of the host lattices containing the 2,2’-pySSpy guest moieties for 3 and 5 were determined as 13.1 and 16.1 A, respectively.4,4‘-PySSpy also afforded the intercal- ationo compound 4 with V205, with an interlayer spacing of 13.3 A. However, we were unable to obtain 4,4‘-pySSpy-VOPO, intercalation compounds. Electronic states of the guest molecules and their arrangements in the interlayer spaces Fig. 3 illustrates the XPS bands of the nitrogen 1sl/2 electrons for compounds 1, 5 and 2,2‘-pySSpy. Compound 1 shows a band at 401.5 eV and a very weak shoulder at 399.2 eV, which are ascribed to the protonated pyridyl nitrogen3*q31 and the unprotonated pyridyl nitrogen, respectively. This finding indi- cates that almost all the guest molecules are protonated in the interlayer space.The band due to vanadium 2p3,, electrons appeared at 517.5 eV for 1, which suggests the reduction of the host lattice from vanadium(v) to vanadium(Iv), as seen for other V205 intercalation compounds.16 Compound 2 exhibited a similar XPS band. Thus, the V205 host lattice affords effective oxidative S-S coupling of the guest molecules, accompanied by the protonation of the pyridyl nitrogen atoms. The bands of the nitrogen 1s1/2 electrons of compounds 3 and 4 also appeared at 401.5 eV, ascribable to the protonated pyridyl nitrogen atom, together with very weak shoulders correspond- ing to unprotonated pyridyl nitrogen. On the other hand, the nitrogen 1sl/2 XP spectrum of compound 5 shows a band due to unprotonated pyridyl nitrogen with an appreciably increased intensity.This suggests that the guest molecules consist of rather increased amounts of the unprotonated pyridyl moieties compared with that of the V205 intercalation compound. This difference of relative amounts of N-protonated sites between the V205 and VOPO, intercalation compounds arises from the proximity of the neighbouring oxidizing vanadium(v) sites of the host lattices and the arrangement of the guest molecules in the interlayer spaces. The V205 host has the vanadium(v) sites in closer proximity,32 which leads to more protonated nitrogen sites of the dithiobipyridine compounds. On the other hand, the VOPO, layer has a rather long distance between adjacent vanadium sites,18 which diminishes the second N-protonation of the dithiobipyridine compound.Molecular geometries of mono-and di-protonated 2,2’- pySSpy moieties calculated by the PM3 method are illustrated in Fig. 4. The bulkiness of the diprotonated molecule in the directiqn perpendicular to the long axis has been estimated to be 5.5 A. Since almost all the guest molecules intercalated into the V205 interlayer space are protonated at the pyridyl nitro- gen atoms, based on the XPS results, the long axes of the diprotonated molecules are arranged approximately parallel to the V,05 host layer. This is similar to the arrangement of ferrocenyl compounds in the V205 interlayer space.16 The enlarged interlayer spacipgs of the intercalation compounds 1 and 3 are 3.6 and 4.3 A, respectively, which are somewhat shorter than the estimated bulkiness of the guest molecules.This is similar to the findings that the intercalatiop of the ferrocenyl derivatives (estimated bulkiness, 5.7-6.8 A33) into the swelling V205 interlayer space resvlted in appreciably shorter interlayer expansions of 4.4-5.4 A.153160 The interlayer expansion for compound 1 is shorter by 0.7A than that for compound 3, which may be due to the formation of a geometri- cally somewhat distorted molecule through S-S coupling in the restricted interlayer space. In the 2,2’-pySSpy-VOoPO, intercalation compound 5, the interlayy distance (16.1 A) indicates an iFterlayer expansion of 11.7 A compared with the spacing (4.4 A) of the anhydrous VOPO, lattice.34 The nitrogen XPS bands of the com- pound exhibit larger amounts of monoprotonated guest moiet- ies than of diprotonated ones.Thus, the long-axis direction of the monoprotonated 2,2’-pySSpy molecule is arranged approximately perpendicular to the two-dimensional VOPO, ~__ 405 400 395 binding energylev Fig. 3 N XPS peaks of (a) V~0,(H,0)o~,,(2-C,H4NSS-2’-CsH4N)o.z 1, (b) VOPO4(HzO)o,3( EtOH)o.,( 2-CcjHdNSS-2’- Fig. 4 Geometries of (a) mono-and (b) N,N’-di-protonated C,H,N)O.l, 5 and (42,2’-PYSSPY 2,T-pySSpy molecules calculated by the PM3 method J. Mater. Chem., 1996, 6(11), 1849-1852 1851 sheet. Since the bulkiness of th,e molecule in the long-axis direction is estimated to be 8.5 A, another arrangement with molecular bilayers is also possible.14 15 16 S. Okuno and G. Matsubayashi, Chem. Lett., 1993,799. P. Aldebert and V. P. Boncour, Mater. Res. Bull., 1983,18, 1263. S. Okuno and G. Matsubayashi, Bull. Chem. SOC. Jpn., 1993, 66, 459. The authors are sincerely grateful to Professor Hiroshi Yoneyama, Osaka University, for the use of the powder X-ray 17 18 J. W. Johnson, A. J. Jacobson, J. F. Brody and S. M. Rich, Inorg. Chem., 1982,21,3820. E. Ruiz-Hitzky and B. Casal, J. Chem. SOC., Faraday Trans., 1986, diffractometer and Mr. Isao Kawafune, Osaka Municipal Technical Research Institute, for the XPS measurements. This research was partly supported by a Grand-in-aid for Scientific Research No. 06640721 from the Ministry of Education, Science and Culture. 19 20 21 82, 1597.M. G. Kanatzidis, C-G. Wu, H. 0. Marcy and C. R. Kannewurf, J. Am. Chem. SOC., 1989,111,4139. Y-J. Liu, D. C. DeGroot, J. L. Schindler, C. R. Kannewurf and M. G. Kanatzidis, J. Chem. SOC., Chem. Commun., 1993,593. H. Nakajima and G. Matsubayashi, Chem. Lett., 1993, 423; J. Mater. Chem., 1995,5, 105. References 22 M. G. Kanatzidis, C-G. Wu, H. 0. Marcy, D. C. Degroot and C. R. Kannewurf, Chem. Mater., 1990,2,222. 1 2 3 4 5 6 7 8 9 10 11 12 Intercalation Chemistry, ed. M. S. Whittingham and A. J. Jacobson, Academic Press, New York, 1982. Inclusion Compounds, ed. J. J. Atwood, J. E. D. Davis and D. D. MacNicol, Oxford University Press, Oxford, 1991, vol. 5. J. W. Johnson, A. J. Jacobson, J. F. Brody and S. M. Rich, Inorg. Chem., 1982,21,3820. K. Bemeke and G.Lagaly, Inorg. Chem., 1983,22,1503. L. Benes, J. Votinsky, J. Kalousova and J. Klikorka, Inorg. Chim. Acta, 1986, 114,47. L. Benes, R. Hyklova, J. Kolousova and J. Votinsky, Inorg. Chim. Acta, 1990,177,71. A. Bouhaouss and P. Aldebert, Mater. Res. Bull., 1983,18, 1248. H. Van Damme, M. Letellier, D. Tinet, B. Kihal and R. Erre, Mater. Res. Bull., 1984,19, 1635. T. Nakato, I. Kato, K. Kuroda and C. Kato, J. Colloid Interface Sci., 1989, 133,447. E. Rodoeiquez-Castellon, A. Jimez-Lopez, M. Martinez-Lala and L. Moreno-Real, J. Inclusion Phenom., 1987,5, 335. G. Matsubayashi and S. Ohta, Chem. Lett., 1990,787. G. Matsubayashi, S. Ohta and S. Okuno, Inorg. Chim. Acta, 1991, 184,47. 23 24 25 26 27 28 29 30 31 32 33 34 G. Matsubayashi and H. Nakajima, Chem. Lett., 1993,31.H. Nakajima and G. Matsubayashi, J. Mater. Chem., 1994,4,1325. H. Nakajima and G. Matsubayashi, unpublished results. J. Lemerle, I. Nejem and J. Lefebvre, J. Inorg. Nucl. Chem., 1980, 42, 17. J. J. P. Stewart, J. Comput. Chem., 1989,10,209,221. J. J. P. Stewart, QCPE Bull., 1989, 9, 10: Y. Inoue, JCPE Newsletter, 1991, 3, 49: T. Takagi, K. Matsumura, A. Noda, N. Onozawa and H. Fujiwara, Bull. Computation Center, Osaka Univ., 1992,22,1: T. Takagi, K. Matsumura, A. Noda, N. Onozawa and H. Fujisawa.Bil1. Computation Centre, Osaka Univ., 1992,22, 1. L. Abello, E. Husson, Y. Repelin and G. Lueazeau, J. Solid State Chem., 1985,56,379. C. Ferragina, M. A. Massucci and G. Mattogno, J. Inclusion Phenom., 1989,7,529. R. B. Borade and A. Clearfield, J. Phys. Chem., 1992,96,6729. R. Ramirez, B. Casal, L. Utrera and E. Ruiz-Hitzky, J. Phys. Chem., 1990,94,8960. M. B. Dines, Science, 1975,188, 1210. B. Jordan and C. Calvo, Can. J. Chem., 1973, 51, 2621; R. Gopal and C. Calvo, J. Solid State Chem., 1971,5,432. 13 S. Okuno and G. Matsubayashi, J. Chem. SOC., Dalton Trans., 1992,2441. Paper 6/02013E; Received 22nd March, 1996 1852 J. Mater. Chem., 1996, 6(11), 1849-1852

ISSN:0959-9428    

DOI:10.1039/JM9960601849

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

MATERIALS CHEMISTRY COMMUNICATIONS Solvate-switchable powder second harmonic generation in a push-pull quinonoid system M. Ravi,"D. Narayana Rao,b Shmuel Cohen,' Israel Agranatd and T. P. Radhakrishnan"" "School of Chemistry and bSchool of Physics, University of Hyderabad, Hyderabad 500 046, India 'Department of Inorganic Chemistry and dDepartment of Organic Chemistry, The Hebrew University of Jerusalem, Jerusalem, 91904, Israel Strong solvate-switchable second harmonic generation has been observed in powders of 7,7-dipyrrolidino-8,8- dicyanoquinodimet hane. Organic molecules and polymers with n-electron conjugation are proving to be a prolific source for the development of non- linear optical (NLO) materials.' Generally, donor-acceptor substituted n-conjugated molecules assembled in a non-centro- symmetric array show large second-order susceptibilities.2 Electric field poling in polymers, oriented Langmuir-Blodgett (LB) film growth and crystal engineering utilising hydrogen bonding, molecular chirality, etc.are standard techniques for the induction of the non-centrosymmetry essential for the observation of quadratic NLO effects such as second harmonic generation (SHG). Several instances of apparently centrosymmetric systems giving rise to SHG have been reported. Weak SHG found3 in powders of aromatic charge-transfer molecules with centrosym- metric crystal structures has been attributed to minor devi- ations from a strictly centrosymmetric lattice. Strong SHG from LB monolayers of centrosymmetric squaraine dyes4 indi- cates that these molecules which form centrosymmetric crystals assemble non-centrosymmetrically in the LB film.Occlusion of chiral molecules has been shown to induce non-centrosym- metry in centrosymmetric lattices and to lead to strong SHG activity.' SHG has been observed in centrosymmetric crystals of racemic mixtures when circularly polarised light was used.6 All these cases involve induction of non-centrosymmetry in centrosymmetric systems in rather unusual ways. A more obvious, though less directed, approach has been to search through solvent-based polymorphs. Polar solvents have been reported to encourage non-centrosymmetric crystal str~ctures.~ Completely SHG-inactive and moderately to strongly SHG- active polymorphs have been grown from different solvents.' An extreme case of solvent-dependent polymorphism has been reported in stilbene and diphenylacetylene derivatives' with one derivative having seven polymorphs with SHG activity ranging from 0.15 to 300 U (1U is the intensity of the SHG signal from urea).Recently, we have observed a novel and interesting case of the role of solvate molecules in determining the SHG capability of polymorphic forms in a quinonoid molecular material. Diamino-substituted dicyanoquinodimethanes, first syn-thesised by a du Pont grouplo in 1962, have large molecular hyperpolarisabilities," and we have shown that the introduc- tion of chiral groups leads to materials capable of moderate to strong SHG.12 Crystals of 7,7-dipyrrolidino-8,8-dicyanoqui-nodimethane (DPDQ) grown from acetonitrile were found to be centrosymmetric and SHG-inactive.The crystals obtained from chloroform are also found to be centrosymmetric; how- ever they lose the solvate molecules upon brief exposure of the crystal to the air, giving rise to powders capable of strong SHG. The SHG activity can be reversibly switched off and on by exposure to chloroform and subsequent drying. Here we present the structures of the crystals from acetonitrile and chloroform and provide an explanation for this novel phenom- enon of solvate switching of the powder SHG in the latter. 9'NCACN DPDQ crystals grown from acetonitrile solution (DPDQ- A).(. belong to P2Jc space group: and contain a water molecule of crystallisation per DPDQ from the solvent, which even upon drying retained traces of moisture.Crystals obtained from a chloroform solution (DPDQ-C)? were immediately coated with grease to prevent solvent loss and crystal degra- dation. Structure analysis indicated a P2Jn space group: and a solvate molecule of chloroform per DPDQ. The molecular geometry of DPDQ is nearly identical in the two crystals. The diaminomethylene unit is twisted out of the six-membered ring plane by ca. 56", a feature common among these molecule^.'^ The bond lengths in the six-membered ring conjugation unit t Satisfactory spectral and analytical data were obtained. $ Crystal data: DPDQ-A (C18H20N4.H20), moyoclinic, space group P~,/c, a=b6.473(2), b=8.481(1), ~=25.383(3) A, p= 105.20(1)", V= 3422.2(6) A3, 2=8, p= 1.21 g anw3,p(Cu-Ka)=5.81 cm-', no.unique reflections=5492, no. reflections (I234 =4200, R =0.046. DPDQ-C (C,,H,,N,~CHCl,), monoc$nic, space group P2,/n, a= !3.278( l), b=17.658(1), ~=8.907(1)A, p=103.11(1)", V=2034.0(8)A3, 2=4, p= 1.34 g cmP3, p(Cu-Ka)=42.40 cm-', no. unique reflections =3071, no. reflections (I22a,) =2296, R =0.091. The crystal structure data were collected on an Enraf-NoQius CAD4 computer-controlled diffractometer. Cu-Ka (A =1.54178 A) radiation with a graphite crystal monochromator in the incident beam was used. The standard CAD4 centring, indexing and data collection programs were used. The unit- cell dimensions were obtained by a least-squares fit of 24 centred reflections in the range 23 GO<31".Atomic coordinates, thermal parameters and bond lengths and angles have been deposited at the Cambridge Crystallographic Data Centre (CCDC). See Information for Authors, J. Mater. Chem., 1996, Issue 1. Any request to the CCDC for this material should quote the full literature citation and the reference number 1145/14. J. Mater. Chem., 1996,6(ll), 1853-1855 1853 indicate strong benzenoid character and hence the zwitterionic nature of this molecule. The resultant dipole moment of this molecule is expected to be high (AM1 calculation indicated ,u = 14.1 D) which probably leads to a preference for centrosym- metric packing in the crystals. Semiempirical (AM 1/MOPAC93) finite-field calculationsi4 gave an appreci- able static hyperpolarisability of -39 x low3'esu.Layers of molecules roughly parallel to the ac plane are seen in the DPDQ-C structure. Projection of these layers in the ab plane indicates zigzag chains of molecules running along the a axis with the adjacent dipoles at an angle of ca. 30" [Fig. 1(a)]. The centrosymmetrically related layer along the b axis (the longest unit-cell axis) is well separated from the first by a layer of chloroform molecules. In DPDQ-A we find a layered structure consisting of stacks of DPDQ, approximately perpen- dicular to the longest unit-cell axis, c; however, there is no distinct layer of solvate molecules [Fig. 1(b)].Inspection of interatomic distances indicates that the water molecules of crystallisation are hydrogen-bonded to the cxano groups (O---Ndistances are found to be 2.73 and 3.10A in the two molecules in the asymmetric unit).DPDQ recrystallised from acetonitrile, tetrahydrofuran, methanol, ethanol and dichloromethane did not show any SHG either in the crystal or in the dried powders. Freshly recrystallised DPDQ-C suspended in chloroform did not show any SHG. Upon drying, the crystals lost their transparency and formed a flaky powder. In a Kurtz-Perryls experiment using 1064 nm radiation from a ns-pulsed Nd : YAG laser, this powder showed a strong SHG of 30-40 U. This SHG activity is highly reproducible provided the sample has been purified by several recrystallisations. We monitored a couple of samples for 15 months and found that the SHG activity remained undiminished.We carried out an experiment in which dry DPDQ-C powder was placed in a narrow tube attached through a stopcock to a flask containing chloroform. While monitoring the SHG activity of the powder the stopcock was opened and the chloroform was warmed gently to allow vapours to flow over the powder. The green light emitted by the sample diminished considerably and vanished completely when the powder was wetted by traces of condensed chloro- form. There was no dissolution since the solubility of DPDQ in chloroform at room temperature is extremely low. After closing the stopcock the sample was dried using a hot air blower whereupon the SHG returned to the original intensity.This experiment could be repeated several times. We carried out control experiments with other compounds (having similar SHG activity and solubility characteristics to those of DPDQ) that we have prepared,12 but no reduction of SHG activity on exposure to chloroform vapours or wetting by chloroform could be detected. The structure of DPDQ-C described above provides an explanation for this interesting phenomenon. The chloroform 0 .. 9 "$, 0 0 Fig. 1 Unit-cell view of (a) DPDQ-C along the c axis and (b) DPDQ-A along the b axis; the C atoms of the pyrrolidine rings and the H atoms are omitted for clarity and the N atoms of the pyrrolidine rings are shown as filled circles. 1854 J. Muter. Chem., 1996,6(11), 1853-1855 layer separates layers of DPDQ molecules related by the centre of inversion.Loss of the chloroform layer upon exposure or drying of the crystals (elemental analysis of the dried powders shows the complete absence of chloroform) results in flaky powders, wherein the original structure has collapsed to intrin- sically non-centric layers of DPDQ which are no longer related by a centre of inversion. In DPDQ-A the water molecules could be partially removed by drying under vacuum (elemental analysis indicated 0.4 mol of water of crystallisation) and driven out completely by extended drying under vacuum at 150°C (confirmed by the disapperance of the O-H stretching absorp- tion in the IR spectrum of such samples). These samples showed no SHG indicating that the type of layer separation seen in DPDQ-C does not occur here.The mechanism of reentry of chloroform into DPDQ-C is not clear at this point; however, it obviously resurrects a centric structure so that the SHG activity is lost. We have attempted to prepare single crystals of DPDQ by solvent-free methods such as sublimation to gain insight into the unsolvated crystal structure. However, these attempts were unsuccessful since the compound decom- poses just before melting (mp 290 "C). The solvate loss provides a novel mode of induction of non- centrosymmetry in a centric structure. The reentry of the solvate and reversibility of this process is unique and the associated switching of the SHG activity in DPDQ-C could perhaps be exploited for sensor applications.Currently we are investigating suitable host materials for the inclusion of the DPDQ-C powder for more detailed studies of the sensor efficiency. M. R. thanks the CSIR, New Delhi for the senior research fellowship and T. P. R. and D. N. R. thank the DST, New Delhi for financial support. References 1 W. Nie, Adv. Muter., 1993,5, 520. 2 Chem. Rev., 1994,94. 3 I. Ledoux, J. Zyss, J. S. Siegel, J. Brienne and J.-M. Lehn, Chem. Phys. Lett., 1990,172,440. 4 G. J. Ashwell, G. Jefferies, D. G. Hamilton, D. E. Lynch, M. P. S. Roberts, G. S. Bahra and C. R. Brown, Nature (London), 1995,375,385. 5 I. Weissbuch, M. Lahav, L. Leiscrowitz, G. R. Meredith and H. Vanherzeele, Chem. Muter., 1989,1, 114.6 E. W. Meijer, E. E. Havinga and G. L. J. A. Rikken, Phys. Rev. Lett., 1990, 65, 37; E. W. Meijer and E. E. Havinga, Synth. Met., 1993,55-57,4010. 7 H. Tabei, T. Kurihara and T. Kaino, Appl. Phys. Lett., 1987, 50, 1855. 8 T. Nogami, H. Nakano, Y. Shirota, S. Umegaki, Y. Shimizu, T. Uemiya and N. Yasuda, in Nonlinear Optics of Organics and Semiconductors, ed. T. Kobayashi, Springer Proc. Phys. vol. 36, Springer-Verlag, Berlin, 1989, p. 232; K. Huang, D. Britton, M. C. Etter and S. R. Byrn, J. Mater. Chem., 1995,5,379. 9 Y. Wang, W. Tam, S. H. Stevenson, R. A. Clement and J. Calabrese, Chem. Phys. Lett., 1988,148,136. 10 L. R. Hertler, H. D. Hartzler, D. S. Acker and R. E. Benson, J. Am. Chem. SOC.,1962,84,3387. 11 S. J. Lalama, K. D. Singer, A. F. Garito and K. N. Desai, Appl. Phys. Lett., 1981, 39, 940; M. Ravi and T. P. Radhakrishnan, J. Phys. Chem., 1995,99,17624. 12 M. Ravi, D. Narayana Rao, S. Cohen, I. Agranat and T. P. Radhakrishnan, Curr. Sci. (India), 1995, 68, 1119; M. Ravi, D. Narayana Rao, S. Cohen, I. Agranat and T. P. Radhakrishnan, J. Mater. Chem., 1996,6, 1119. 13 M. Ravi, S. Cohen, I. Agranat and T. P. Radhakrishnan, Struct. Chem., 1996,7,225. 14 H. A. Kurtz, J. J. P. Stewart and K. M. Dieter, J. Comput. Chem., 1990,11,82. 15 S. K. Kurtz and T. T. Perry, J. Appl. Phys., 1968,39, 3798. Paper 6104976A; Received 16th July, 1996 J. Mater. Chem., 1996, 6(11), 1853-1855 1855

ISSN:0959-9428    

DOI:10.1039/JM9960601853

出版商:RSC    

年代:1996

数据来源: RSC

摘要:

Synthesis of Al-intercalated montmorillonites using microwave irradiation G. Fetter,"?"G. Heredia,"A. M. Maubert" and P. Boschb "Universidad Autbnoma Metropolitana-Azcapotzalco, Divisibn de C.B.I., Area de Quimica Aplicada, Avenida Sun Pablo 180, C.P. 2200, Mixico D.F., Mexico bUniversidad Autbnoma Metropolitana-Iztapalapa,Departamento de Quimica, Area de Catalisis, Michoachn esquina Purisima, C.P. 09340, Mixico D. F., Mexico Microwave irradiation has been used to synthesize intercalated clays during the intercalation reaction. The obtained montmorillonites, when compared to conventionally prepared clays, presented a higher surface area (340 m2 g cJ: 244 m2 g-'). This method presents the advantage of reducing the synthesis time and the possibility of handling highly concentrated clay suspensions and intercalating solutions. The potential use of pillared clays (PILCs) as catalysts or sorbents is already well Since the pioneering works of Vaughan et uL3 on aluminium PILCs and Yamanaka and Brindley4 on zirconium PILCs, many natural and synthetic smectites have been pillared with diverse pillaring agents.Metallic complex polycations are nowadays preferred as they provide thermally stable PILCs with required large pore diameters. The use of different pillaring agents is the most studied theme in pillared clays. However, in all these syntheses, a main disadvantage is the very long time required for intercalation with samples being treated from 3 up to 20 hours.' Recently, the use of microwave radiation has been reported during intercalation5 or the drying of samples.6 In the first case, the iron oxide pillared clays were characterized by 57Fe Mossbauer spectroscopy and it was found that the samples were prepared in <1 h.Nevertheless, this successful result has not been applied to the synthesis of clays pillared by conventional hydroxy cations of aluminium or zirconium. The microwave drying of boehmite sol interca- lated smectites provided samples with identical thermal and electrical properties to those prepared by conventional means. However, the microwave-dried samples had a distinctly higher surface area of 120 m2 g-', compared with 94 m2 g-' for the oven-dried sample. Similarly, there was a clear difference in the morphological features of the two samples, the air-dried sample had a close packed structure while the microwaved one was delaminated and porous.It seems that no other workers have studied the applications of microwave irradiation to the synthesis of pillared clays. In this work, we present the synthesis of aluminium interca- lated montmorillonite in the presence of microwave radiation, and the effect of irradiation time on the structure as well as surface area is discussed. The materials are also compared with a conventionally prepared clay treated for 18 h at room temperature. A commercial Na-montmorillonite (Bentonite Sigma) was studied and used as received, i.e. it was not purified and showed low contents of quartz and organic compounds as determined by X-ray diffraction and IR spectroscopy respect- ively.The intercalating aluminium solution was aluminium t Present address: Instituto PolitCcnico Nacional, ESIQIE, SEPI. UPALM, Edif. 8, 3er. piso. C.P. 07738 MCxico, D.F., Mexico. chlorohydrate [Chlorhydrol, Reheis Chemical Co.; 50% (m/m) Al,O,, OH/Al =2.01. The aluminium solution was 2.5 mol dm-, with respect to A1 and it was used without dilution. In a test tube, 3 g of montmorillonite were dispersed in 30ml of deionized water (10 mass% clay concentration) and manually agitated and 4 g of aluminium chlorohydrate (5 mmol A1 per gram of clay) were added to the montmorillon- ite slurry. The glass tube was then hermetically closed and placed in a commercial microwave oven (Sharp Carousel, model R-5456M) operating at 2.45 GHz and at a power level of 90 W for 2, 3, 5, 7, 10 or 15 min.The product was then washed by redispersing it in water and separated by decan- tation. This procedure was repeated until the supernatant was free of chloride ions as determined by AgNO,. These samples are designed M-A1-2, M-A1-3, M-A1-5 where the numbers define the microwave irradiation time in min. For comparison purposes, the Sigma clay was also inter- calated following a conventional method: 5 g of clay were dispersed in 1000ml of deionized water, (i.e. 0.5 mass% concentration) and stirred for 1 h, and 5.3 g of aluminium chlorohydrate diluted in 78 ml of deionised water was added dropwise to the clay slurry. The mixture was then stirred for 18 h. The product was washed as for the microwave samples and, as for the other samples, was not calcined; this sample is designated C-Al.X-Ray diffraction patterns were obtained with a Siemens D-500 diffractometer and an X-ray tube with a copper anode. A diffracted beam monochromator selected the Ka radiation. The BET (N,) surface areas were determined with Micromeritics ASAP 2000 equipment on samples outgassed at 573 K for 5 h. IR spectra were obtained in a FTIR Nicolet 750 instrument, analysing a 1 mass% sample in a KBr wafer. Fig. 1 shows the X-ray diffraction patterns for 28=2-15" of the various samples. In the intercalate< clays, the interlayer distance dool did not yary and was 19 A, Table 1. This value is similar to the 19.7 A found in the C-A1 preparation, within experimental error.It seems therefore that the structural parameters of the resulting samples are similar and independent of the preparation method. Furthermore, as the layer separa- tion is the same, the intercalated oligomer is expected to be [Al,,O,(OH),, -12H,0I7+ (ref. 7). Fig. 2 compares the IR spectra of the original clay, the M- A1 and C-A1 samples. All spectra show the same well known Table 1 Microwave irradiation time effect on the (001) interlayer spacing sample do01 M-A1-2 18.9 M-A1-5 18.9 M-Al-7 19.1 M-Al-10 19.0 M- Al- 15 18.8 C-A1 19.7 a k0.5 A. J. Muter. Chem., 1996, 6(11), 1857-1858 1857 1 6.00 10.00 14.00 2Bldeg rees Fig. 1 X-Ray diffraction patterns of the samples: C-A1 (a), M-A1-2 (b), M-Al-5 (c),M-A1-7 (d), M-Al-10 (e) and M-A1-15 (f) I t I I 4000 3000 2000 1000 v1crn-l Fig.2 IR spectra of the initial clay (a) and intercalated samples: C-A1 (b),M-Al-2 (c),M-Al-5 (d), M-Al-7 (e),M-Al-10 (f)and M-A1-15 (g) bands, but they differ in the region 2800-2600cm-'. Two small absorption bands observed in the original clay and in the C-A1 sample are absent in the M-A1 samples. These bands are attributed to organic impurities,' which are eliminated during microwave irradiation. This treatment may result in high enough temperatures to decompose the organic materials, 1858 J. Mater. Chem., 1996, 6(ll), 1857-1858 itE 200 zi 100 "I I ..I. 0 2 4 6 8 10 12 14 16 timdmin Fig.3 Microwave irradiation time effect on intercalated clay surface areas but not high enough to alter the structure of the intercalated species as the layer separation is the same. A surface area of 293-347m2 g-' is observed in the M-A1 samples (Fig. 3). This value is 20-30% higher than that measured for the C-A1 sample (244 m2 g-I). As the surface area found for all the M-A1 samples is similar, within experimental error, in these preparations the micro- porosity, determined by the number and distribution of intercalated species into the clay particles, seems to be homogeneous. To be sure that the results obtained with this method were reproducible, several samples were prepared twice. For instance the first and second M-A1-5 samples exhibited do,, interlayer spacings and specific surface areas of 19.2, 18.9A and 347, 330 m2 g-', respectively.It seems, therefore, that the interca- lated clays show better reproducibility using this new method. Another advantage, clearly shown by our results, is the possibil- ity to synthesize reproducible intercalated clays from micro- wave irradiated concentrated clay suspensions and pillaring solutions instead of conventionally diluted suspensions and solutions. In conclusion, microwave irradiation during intercalation offers, against conventional preparation, the following advan- tages: (i) reduced intercalation time, (ii) the surface area is independent of the irradiation time (an intercalated clay with a surface area as high as 347 m2 g-' was obtained after 5 min irradiation time); (iii) organic impurities present in the original clay are eliminated and (iv) highly concentrated clay suspen- sions and aluminium chlorohydrate solutions can be employed.Since microwaved intercalated samples (M-Al) showed inter- layer distances comparable to the C-A1 sample, the nature of the pillaring agents is assumed to be the same, uiz. [A11304(0H)2412H20I7'. G. F. thanks CONACYT for a fellowship (Catedra Patrimonial) which provided the personal support to carry out this work. References 1 F. Figueras, Catal. Rev. Sci. Eng., 1988,30,457. 2 R. Burch and C. I. Warburton, J. Catal., 1986,97,511. 3 D. E. W. Vaughan, R. J. Lussier and J. S. Magee, US Pat., 4 176 090, 1979; 4 248 739,1981. 4 S.Yamanaka and G. W. Brindley, Clays Clay Miner., 1978,26,21. 5 F. J. Berry, K. K. Rao and G. Oates, Hyperfine Interact., 1994, 83,343. 6 K.G.K. Warrier, P. Mukundan, S. K. Ghosh, S. Sivakumar and A. D. Damodaran, J. Muter. Sci., 1994,29, 3415. 7 D. Plee, F. Borg, L. Gatineau and J. J. Fripiat, J. Am. Chem. SOC., 1985,107,2362. 8 M. M. Mortland and V. Berkheiser, Clays Clay Miner., 1976,24,60. Communication 6/05 138C; Received 23rd July, 1996

ISSN:0959-9428    

DOI:10.1039/JM9960601857

出版商:RSC    

年代:1996

数据来源: RSC