Solid-State 51V NMR Investigation of the Intercalation of Alkylamines

Mar 31, 2010 - For the intercalation with short chain amines (propylamine, pentylamine, and hexylamine), the C−C chain of the amines is tilted with ...
0 downloads 0 Views 2MB Size
Download PDF
pubs.acs.org/Langmuir © 2010 American Chemical Society

Solid-State 51V NMR Investigation of the Intercalation of Alkylamines into Layered r-Vanadyl Phosphate Jianfeng Zhu and Yining Huang* Department of Chemistry, The University of Western Ontario, London, Ontario, Canada N6A 5B7 Received January 18, 2010. Revised Manuscript Received March 22, 2010 The intercalation behavior of layered R-phase vanadyl phosphate, R-VOPO4 3 2H2O (R-VP), with alkylamine was investigated by 51V solid-state NMR in combination with powder XRD. The XRD results show that the amines form bimolecular layers upon intercalation. For the intercalation with short chain amines (propylamine, pentylamine, and hexylamine), the C-C chain of the amines is tilted with respect to the inorganic basal plane. The amines with a longer alkyl chain (dodecylamine and hexadecylamine) tend to adopt an orientation where the C-C chain direction is perpendicular to the VP layer. For the amine with eight carbon atoms (octylamine), the intercalation results in two coexisting phases with different chain orientations. 51V solid-state NMR was used to directly probe the effect of intercalation on the metal center environments. Both 51V magic-angle spinning and static spectra of R-VP intercalated with different amines were obtained at different magnetic fields, and they are sensitive to intercalation. The intercalation induces the 51V isotropic chemical shift to move toward deshielded direction. 51V chemical shielding parameters such as the span are sensitive to the orientation of the amine chain with respect to the VP basal plane. For the V centers interacting with the amines having a tilted orientation, the 51V span gradually decreases with increasing alkyl chain length. However, the span of the 51V atoms interacting with the amines perpendicular to the VP layer is larger and independent of the length of the alkyl chain. The 51V NMR data indicate that for the R-VPs intercalated with longchain amines, such as dodecylamine and hexadecylamine, the amines can assume both tilted and perpendicular orientations.

Introduction As an important member of the layered metal phosphate family, MP (M = Zr, Ti, V, Nb, Mo, Mg, and Hf),1,2 layered vanadyl phosphates (VPs) have received much attention because of their potential applications in catalysis, lithium ion batteries, and intercalation.3 For example, layered VPs are selectively used to oxidize n-butane or n-butene to maleic anhydride.4-6 For this purpose, a new family of mixed VP-NbP catalysts was prepared several years ago.7 VPs have also shown potential as catalysts for the oxidation of other light hydrocarbons8-10 and alkyl aromatics.11 Furthermore, VPs are considered to be good candidates for new cathode materials in lithium ion rechargeable batteries because of their unique electrochemical behavior.12,13 Among many layered VPs and their derivatives, R-VOPO4 3 2H2O (R-VP)14,15 has been extensively examined. The structure of *Corresponding author. E-mail: [email protected]. Tel: 519-661-2111, ext 86384. Fax 519-661-3022. (1) Clearfield, A.; Costantino, U. Comp. Supra. Chem. 1996, V7, 107 and references therein. (2) Auerbach, S. M.; Carrado, K. A.; Dutta, R. K. Handbook of Layered Materials; Marcel Dekker: New York, 2004; pp 313-371 and references therein. (3) Centi, G., ed; Forum on Vanadyl Pyrophosphate. Catal. Today 1994, 16, 1-153. (4) Ait-Lachgar, K.; Tuel, A.; Brun, M.; Abon, M. J. Catal. 1998, 177, 224. (5) Taufig-Yap, Y. H.; Goh, C. K.; Bartley, J. Catal. Lett. 2009, 130, 327. (6) Frey, J.; Ooi, Y. S.; Hunger, M. Solid State Nucl. Magn. Reson. 2009, 35, 130. (7) Duarte de Farias, A. M.; Volta, J.-C. J. Catal. 2002, 208, 238. (8) Lisi, L; Patrono, P.; Ruoppolo, G. J. Mol. Catal. A: Chem. 2003, 42, 609. (9) Conte, M.; Budroni, G.; Bartley, J. K.; Hutchings, G. J. Science 2006, 313, 1270. (10) Benser, E.; Glaum, R.; Dross, T.; Hibst, H. Chem. Mater. 2007, 19, 4341. (11) Marin, A.; Lucke, B. Catal. Today 1996, 32, 279. (12) Dupre, N.; Gaubisher, J.; Quarton, M. Solid State Ionics 2001, 140, 209. (13) Akimura, Y.; Goward, G. R.; Nazar, L. F. Chem. Mater. 2008, 20, 4240. (14) Ladwig, G. Z. Anorg. Allg. Chem. 1965, 338, 266. (15) Bordes, E.; Courtine, P.; Pannetier, G. Ann. Chim. 1973, 8, 105. (16) Tietze, H. R. Aust. J. Chem. 1981, 34, 2035.

Langmuir 2010, 26(12), 10115–10121

R-VP (Figure 1) was determined in 1981 by Tietze16 using singlcrystal X-ray diffraction (XRD). The layers of R-VP contain distorted VO6 octahedra. The V atom in each VO6 octahedron is connected to four PO4 tetrahedra in the equatorial plane via bridging oxygen atoms. The two axial groups are (1) an oxygen atom strongly bound to the vanadium via a V=O double bond and (2) a weakly coordinated water molecule, with the former having a shorter (1.567 A˚) and the latter having a longer V-O (2.233 A˚) distance than the equatorial V-O bonds (1.908 A˚). The layers are electrically neutral and held together by hydrogen bonds between the coordinated water molecules and the water molecules occluded in the interlayer space. The basal spacing of R-VP is 7.4 A˚ as determined from the crystal structure. Because the layers are held together by the weak bonding interactions, guest species readily intercalate into the interlayer galleries. The intercalations of varieties of organic compounds17 such as amines18-21 and alcohols22 into R-VP have been studied using mainly thermogravimetric analysis (TGA) and IR spectroscopy. The detailed structural information on intercalated VP materials is important in understanding the intercalation process and in designing new composite materials for specific applications. However, the crystal structure of intercalated layered VPs is rarely available because, as is the case for many MPs, it is usually very difficult to grow high-quality single crystals suitable for X-ray diffraction. (17) Kalousova, J.; Votinsky, J.; Benes, L.; Melanova, K.; Zima, V. Collect. Czech. Chem. Commun. 1998, 63, 1. (18) Beneke, K.; Lagaly, G. Inorg. Chem. 1983, 22, 1503. (19) Morris, M.; Dyer, A.; McCabe, R. W. J. Mater. Chem. 1995, 5, 1427. (20) Dasgupta, S.; Agarwal, M.; Datta, A. J. Mol. Catal. A: Chem. 2004, 223, 167. (21) Datta, A.; Dasgupta, S.; Agarwal, M.; Ray, S. S. Micoporous Mesoporous Mater. 2005, 83, 114. (22) Benes, L.; Zima, V; Melanova, K. J. Inclusion Phenom. Macrocyclic Chem. 2001, 40, 131.

Published on Web 03/31/2010

DOI: 10.1021/la100232s

10115

Article

Zhu and Huang

Figure 1. Structure of R-VOPO4 3 2H2O: (A) the layered structure and (B) the detailed coordination in each layer. The hydrogens are omitted for clarity.

Furthermore, the intercalation of guest species often results in a powder sample with poor crystallinity, which prevents one from accurately solving the structure from powder XRD data. Solid-state NMR is a complementary technique to X-ray diffraction. However, only a few 31P and 13C MAS NMR studies4,19-25 have been performed on intercalated VPs. The effect of intercalation on the metal center local environment has not been directly probed by solid-state 51V NMR. To our knowledge, 51V magic-angle spinning (MAS) NMR has been applied only to study the dehydrated phases of R-VP.6,26-28 This is somewhat surprising because 51V is an NMR-favorable nucleus. 51V (I=7/2) has a relatively high gyromagnetic ratio (γ) comparable to that of 23 Na, a high natural abundance (99.76%), and a relatively small quadrupole moment.29 The high sensitivity of 51V makes 51V solid-state NMR a powerful technique to use in characterizing vanadium oxide materials and V-containing catalysts,30,31 V-implicating biological and medical systems,32-39 vanadia gels40 and systems with paramagnetic centers.41 The reported NMR parameters show that 51V usually exhibits quadrupolar coupling (23) Bartley, J. K.; Kiely, C. J.; Wells, R. P.; Hutchings, G. J. Catal. Lett. 2001, 72, 99. (24) Sananes, M. T.; Tuel, A.; Volta, J. C. J. Catal. 1994, 145, 251. (25) Sananes, M. T.; Tuel, A.; Hutchings, G. J.; Volta, J. C. J. Catal. 1994, 148, 395. (26) Siegel, R.; Dupre, N.; Quarton, M.; Hirschinger, J. Magn. Reson. Chem. 2004, 42, 1022. (27) Lapina, O. B.; Khabibulin, D. F.; Shubin, A. A.; Bondareva, V. M. J. Mol. Catal. A: Chem. 2000, 162, 381. (28) Meisel, M.; Wolf, G.-U.; Worzala, H.; Eichele, K.; Grimmer, A.-R. Phosphorus Res. Bull. 2000, 11, 81. (29) Pyykko, P. Mol. Phys. 2001, 99, 1617. (30) Lapina, O. B.; Mastikhin, V. M.; Shubin, A. A.; Krasilnikov, V. K.; Zamaraev, K. I. Prog. NMR Spectrosc. 1992, 24, 457. (31) Lapina, O. B.; Khabibulin, D. F.; Shubin, A. A.; Terskikh, V. V. Prog. Nucl. Magn. Reson. Spectrosc. 2008, 53, 128 and references therein. (32) Rehder, D.; Polenova, T.; Buehl, M. Ann. Rep. NMR Spectrosc. 2007, 62, 49. (33) Pooransingh-Margolis, N.; Renirie, R.; Hasan, Z.; Wever, R.; Vega, A. J.; Polenova, T. J. Am. Chem. Soc. 2006, 128, 5190. (34) Schweitzer, A.; Gutmann, T.; Wachtler, M.; Breitzke, H.; Buchholz, A.; Plass, W.; Buntkowsky, G. Solid State Nucl. Magn. Reson. 2008, 34, 52. (35) Ooms, K. J.; Bolte, S. E.; Baruah, B.; Choudhary, M. A.; Crans, D. C.; Polenova, T. Dalton Trans. 2009, 3262. (36) Pizzala, H.; Caldarelli, S.; Eon, J.; Rossi, A. M.; Laurencin, D.; Smith, M. E. J. Am. Chem. Soc. 2009, 131, 5145. (37) Ooms, K. J.; Bolte, S. E.; Smee, J. J.; Baruah, B.; Crans, D. C.; Polenova, T. Inorg. Chem. 2007, 46, 9285. (38) Bolts, S. E.; Ooms, K. J.; Polenova, T.; Baruah, B.; Crans, D. C.; Smee, J. J. J. Chem. Phys. 2008, 128, 052317. (39) Smee, J. J.; Epps, J. A.; Ooms, K.; Bolte, S. E.; Polenova, T.; Baruah, B.; Yang, L.; Ding, W.; Li, M.; Willsky, G. R.; la Cour, A.; Anderson, O. P.; Crans, D. C. J. Inorg. Biochem. 2009, 103, 575. (40) Fontenot, C. J.; Wiench, J. W.; Pruski, M.; Schrader, G. L. J. Phys. Chem. B 2001, 105, 10496. (41) Shubin, A. A.; Khabibulin, D. F.; Lapina, O. B. Catal. Today 2009, 142, 220.

10116 DOI: 10.1021/la100232s

constants (CQs) ranging from 2 to 6 MHz and that the chemical shift (CS) interaction usually makes a significant contribution to the spectra with chemical shift anisotropy (CSA) values from several hundreds to thousands of parts per million.30-32 The presence of both quadrupolar interaction and CSA can considerably complicate 51V NMR spectra and their analysis. In addition to the central (1/2 T -1/2) transition, more often than not,42 the satellite (mI=(1/2 T (3/2, (3/2 T (5/2, (5/2 T (7/2) transitions are also present in the spectra because of the usually small CQ, which further complicates the spectral analysis. Therefore, theoretical calculations34,35,43,44 were often performed to aid in the analysis of 51V NMR data. In the present study, the local environments of the vanadium centers in R-VP and its alkylamine intercalated derivatives were investigated using solid-state 51V static and MAS NMR. 51V NMR parameters including the isotropic chemical shift (δiso), span (Ω), skew (κ), quadrupolar coupling constant (CQ), and electric field gradient (EFG) tensor asymmetry parameter (ηQ) were obtained. To obtain reliable NMR parameters, 51V spectra were acquired at three field strengths. The effect of the intercalation of alkylamines on the local structures of vanadium was discussed.

Experimental Section R-VP was synthesized using procedures described by Hutchings23 and Hirschinger.26 Typically, 5.0 g of V2O5 was refluxed in a mixture of 27.4 mL of H3PO4 (85%) and 80 mL of water for 24 h. The yellow solids were recovered by vacuum filtration. The product was washed with water and acetone and then dried under ambient conditions. The alkylamine-intercalated R-VPs were prepared by simply stirring 0.2 g of R-VP with approximately 20 mL of pure amines (propylamine, pentylamine, hexylamine, and octylamine) or amine/ethanol solutions (dodecylamine and hexadecylamine) for 1 to 2 days. The color of the solids gradually changed from yellow to white. The solid product was recovered by centrifugation, washed with ethanol, and finally dried in air at room temperature. The identity of the synthesized materials was confirmed by comparing the powder XRD patterns with those reported in the literature.17-21,23,26 Powder XRD measurements were performed on a Rigaku rotating anode diffractometer (45 kV/160 mA) using graphite-monochromated Co KR radiation with a wavelength of 1.7902 A˚. The powder XRD patterns were recorded within the range of 5 e 2θ e 65 with a 10/min step width and a 6 min (42) Nakashima, T. T.; Teymoori, R.; Wasylishen, R. E. Magn. Reson. Chem. 2009, 47, 465. (43) Lo, A. Y. H.; Hanna, J. V.; Schurko, R. W. Appl. Magn. Reson. 2007, 32, 691. (44) Gee, B. A. Solid State Nucl. Magn. Reson. 2006, 30, 171.

Langmuir 2010, 26(12), 10115–10121

Zhu and Huang count time. For long-alkyl-chain amine-intercalated VPs, a larger 2θ range of 2-70 was used. 51 V NMR experiments were performed at three different fields: 9.4 (105.15 MHz), 14.1 (157.66 MHz), and 21.1 T (236.78 MHz). The experiments at 9.4 T were performed on a Varian Infinityplus 400 WB spectrometer using a 5 mm MAS probe. A nonselective π /2 pulse width of 4.0 μs, corresponding to a selective π/2 pulse width of 1.0 μs for the 51V central transition, was calibrated on VOCl3 liquid. A recycle delay of 0.5 s was applied. To determine the isotropic shift, MAS spectra at two spinning rates, 7 and 10 kHz, were obtained. 51V static spectra were acquired using the spin echo sequence with the phase cycling developed by Oldfield et al.,45 and a π/2 refocusing pulse (π/2 - τ1 - π/2 - τ2 - acq, where τ represents interpulse delays, τ1 =50 μs, and τ2 =40 μs) was used. The spectra at 14.1 T were acquired on an Inova 600 spectrometer using a 3.2 mm MAS probe. The selective π/2 pulse width was 1.0 μs, and the recycle delay was 1.0 s. The MAS spectra were acquired at two spinning rates of 10 and 13 kHz. The interpulse delays were the same as those used at 9.4 T. The experiments at 21.1 T were performed on a 900 MHz Bruker Avance II spectrometer at the National Ultrahigh-field NMR Facility for Solids in Ottawa, Canada. The MAS spectra were acquired using a 2.5 mm MAS probe at two spinning rates of 20 and 30 kHz. A selective π/2 pulse width of 1.3 μs and a recycle delay of 1.0 s were applied. The static echo spectra were obtained using a 4.0 mm MAS probe. The selective π/2 pulse width was 0.75 μs. 51V chemical shifts were referenced to liquid VOCl3 (δiso =0 ppm). The 31P NMR experiments were performed on a Varian Infinityplus 400 WB spectrometer. A 5 mm MAS probe was used, and the samples were spun at 10 kHz. A π/6 pulse (1.5 μs) and a recycle delay of 60 s were applied. 51 V spectra were simulated by taking into account both the central and the satellite transitions. The 51V NMR parameters including CQ, ηQ, δiso, Ω, and κ were determined by analytical simulations of NMR spectra using the WSOLIDS1 software package provided by Prof. R. E. Wasylishen (University of Alberta). The principal components of the EFG tensor (VXX, VYY, and VZZ) are related to quadrupolar coupling constant CQ and asymmetry parameter ηQ according to the following definition: |VZZ| g |VYY| g |VXX|; CQ(Hz)=(eVZZQ/h); ηQ =(VXX VYY)/VZZ, where e is the electric charge; Q is the nuclear quadrupole moment. The magnetic shielding parameters are described using the Herzfeld-Berger convention. The chemical shift tensor is described by three principal components ordered such that δ11 g δ22 g δ33. The isotropic chemical shift is the average of the three chemical shift tensor components (δiso = (δ11 þ δ22 þ δ33)/3). The span is the difference between the mostand the least-shielded component (Ω = δ11 - δ33). The skew describes the shape of the CSA powder pattern and is related to the axial symmetry of the CS tensor [κ=3(δ22 - δiso)/Ω]. Fitting the spectrum of a stationary sample normally requires eight adjustable parameters [δiso, CQ, ηQ, Ω, κ, and three Euler angles (R, β, γ) defining the relative orientations of the EFG and CS tensors]. In general, to increase the accuracy of the simulation, the fast MAS spectrum of the central transition should be first obtained and simulated to obtain δiso, CQ, and ηQ because the chemical shift interaction is averaged by MAS. Using the same δiso, CQ, and ηQ, the static spectrum can then be analyzed by varying only five adjustable parameters (Ω, κ, R, β, and γ). However, in the present case, only δiso was determined accurately from the MAS spectra of the central transitions (CTs). The small 51 V CQ values of the materials examined prevent them from being accurately extracted from the MAS spectra. The small CQs, however, allow the satellite transitions (STs) to be observed and the STs provide additional spectral features. Because these features need to be simulated properly, they help with converging on a unique fit of the spectrum. Furthermore, the spectra of a given material examined in this study were acquired at at least two (45) Kunwar, A. C.; Turner, G. L.; Oldfield, E. J. Magn. Reson. 1986, 69, 124.

Langmuir 2010, 26(12), 10115–10121

Article

Figure 2. Powder XRD patterns of (A) R-VP and (B-G) its alkylamine-intercalated compounds. The number of carbons in the alkyl group of each amine is indicated in each pattern. Symbols * and  in (E) are used to distinguish two coexisting phases.

V NMR spectra of R-VOPO4 3 2H2O: (A) MAS at 9.4 T, (B) MAS spectra expanded to show isotropic shifts, and (C) static echos at 9.4 and 21.1 T.

Figure 3.

51

field strengths. The spectra obtained at an additional field impose another constraint for simulation because quadrupolar and CS interactions have different field dependencies. Simulating the line shape was carried out interactively with particular attention being paid to the key spectral features. The experimental error in each measured parameter was determined by a visual comparison of the experimental spectra with the simulated ones. The parameter of concern was varied bidirectionally starting from the best-fit value and all other parameters were kept constant until noticeable differences between the spectra were observed. The reported errors in Ω, κ, ηQ and the Euler angles (R, β, γ) are twice the value of the deviation from the best fit clearly detected by visual inspection, a practice used in the literature.46 (46) Bryce, D. L.; Sward, G. D. J. Phys. Chem. B 2006, 110, 26461.

DOI: 10.1021/la100232s

10117

Article

Zhu and Huang Table 1. Parameters Used to Fit the 51V Static Echo Spectra of the VPsa

compounds δiso (ppm) Ω (ppm) κ CQ(MHz) ηQ R (deg) β (deg) γ (deg) R-VP -685 ( 10 1281 ( 40 0.93 ( 0.10 2.5 ( 0.5 1.00 ( 0.10 0 ( 40 50 ( 20 0 ( 40 -570 ( 10 960 ( 40 0.75 ( 0.10 3.0 ( 0.5 1.00 ( 0.10 0 ( 40 50 ( 20 0 ( 40 C3-VP -553 ( 10 526 ( 40 0.12 ( 0.10 2.7 ( 0.5 0.80 ( 0.10 30 ( 40 140 ( 20 0 ( 40 C5-VP -550 ( 10 488 ( 40 0.12 ( 0.10 2.7 ( 0.5 0.80 ( 0.10 30 ( 40 140 ( 20 0 ( 40 C6-VP b S1 -551 ( 10 526 ( 40 0.12 ( 0.10 2.7 ( 0.5 0.80 ( 0.10 20 ( 40 140 ( 20 0 ( 40 C8-VP -540 ( 10 1006 ( 40 0.93 ( 0.10 2.7 ( 0.5 0.80 ( 0.10 0 ( 40 50 ( 20 90 ( 20 S2 S1 -555 ( 10 526 ( 40 0.12 ( 0.10 2.7 ( 0.5 0.80 ( 0.10 30 ( 40 140 ( 20 0 ( 40 C12-VPb S2 -536 ( 10 1006 ( 40 0.93 ( 0.10 2.7 ( 0.5 0.80 ( 0.10 0 ( 40 50 ( 20 90 ( 20 C16-VPb a The isotropic shifts were identified from the 51V MAS spectra with different spinning rates. b The intensity of S1 is 56% in C8-VP and 33% in C12-VP and C16-VP.

Results and Discussion Figure 2A shows the powder XRD pattern of R-VP, which is consistent with those reported in the literature.17-21,23,26 A basal spacing of 7.4 A˚ was obtained from the powder XRD pattern, which is in good agreement with the value determined from the crystal structure.16 The 51V MAS spectra of R-VP with two spinning rates acquired at 9.4 T (Figure 3) reveal that the isotropic shift (δiso) is at -685 ppm. The 51V static spectra of RVP at field strengths of 9.4 and 21.1 T were also obtained and are shown in Figure 3. The powder pattern observed at 21.1 T is almost as broad as that obtained at 9.4 T, indicating that the spectra are dominated by the CSA and the contribution from the quadrupolar interaction is small. This is because the CSA is proportional to and the second-order quadrupolar interaction scales inversely with the strength of the magnetic field. Fitting the static spectra at two fields generates the following parameters (Table 1): Ω=1281 ppm, κ=0.93, CQ =2.5 MHz, ηQ = 1.00, and Euler angles (R, β, γ) = (0, 50, 0). The simulated spectra are shown in Figure 3. The very large CSA and the small CQ extracted via simulations confirm the visual observation that the CSA is indeed the major contributor to the observed central transition. The large CSA and the small CQ are consistent with various related VOPO4 phases reported in the literature.26-28 The relative orientation between the EFG and the CSA is such that the largest EFG principal component, VZZ, forms an angle of 50 with respect to the most shielded component, δ33, and VYY aligns with δ22. The powder XRD patterns of the amine-intercalated VPs are given in Figure 2. The values of basal spacing estimated from the XRD patterns are consistent with those reported in the literature,18-21 confirming the identities of these compounds. The reflections of all of the amine-intercalated VPs are sharp, indicating the high crystallinity. The 51V MAS spectra (Figure 4) of propylamine-intercalated R-VP (C3-VP) show a major peak at -570 ppm (δiso) and a weak signal at -600 ppm. The former originates from C3-VP, and the latter is likely due to a small amount of unreacted V2O5 as an impurity. To assist in the analysis, 51V MAS and static spectra of V2O5 were also obtained. The NMR spectra and the simulation results are presented in Supporting Information Figure S1. Overall, for C3-VP the line-shapes of the static echo spectra match well with the spinning sideband patterns in the MAS spectra. Spectral simulations yielded the following CS and EFG tensor parameters of V center in C3-VP: Ω=960 ppm, κ=0.75, CQ = 3.0 MHz, ηQ = 1.00, and Euler angles (R, β, γ) = (0, 50, 0). It should be pointed out that the inclusion of a V2O5 signal (17% intensity) in the simulation did give a slightly better match with the experimentally observed spectra (Figure 4). The intercalation induces a change in the isotropic shift (δiso) from -685 ppm for the parent R-VP to -570 ppm for C3-VP. The inclusion of the amine also results in a smaller CSA. However, the insertion of 10118 DOI: 10.1021/la100232s

51 V static echo (A) and MAS (B) NMR spectra of propylamine (C3)-intercalated R-VOPO4 3 2H2O at fields of 9.4 (left column) and 21.1 T (right column). (C) The expanded MAS spectra to show isotropic shifts. In A, simulation 2 includes the V2O5 signal with an intensity of 17%.

Figure 4.

propylamine did not cause a significant change in the 51V CQ, suggesting that the overall electronic charge distribution is not drastically perturbed. The 51V MAS spectra of pentylamine-intercalated R-VP (C5-VP) show a single resonance at -553 ppm (Figure S2). No signal due to the V2O5 impurity was observed. The static spectra were fitted, leading to the following parameters: Ω=526 ppm, κ=0.12, CQ = 2.7 MHz, ηQ = 0.80, and Euler angles (R, β, γ) = (30, 140, 0). Comparing R-VP with C5-VP (Table 1) reveals a further shift of δiso in the deshielded direction and a much smaller CSA. Again, the CQ of 51V remains almost unchanged. The relative orientation between the EFG and the CSA tensor also changed: VZZ now makes an angle of 130 with δ11, and VYY makes an angle of 30 with δ22. The mechanisms of the intercalation of alkylamines into R-VP have been examined before. The work by Beneke and Lagaly18 indicated that under mild conditions alkylamines are simply occluded between the layers without forming V-N bonds at room temperature. Amine insertion occurs via hydrogen bonding of the amine headgroup with the VdO group and the water molecule coordinated to the V center, but it was also shown that under harsh conditions (such as dehydrating VP at 300 C, followed by reacting with amines at 100 C) the amine can replace coordinated water and bond directly to the V center.47 In the (47) Benes, L.; Hyklova, R.; Kalousova, J.; Votinsky, J. Inorg. Chim. Acta 1990, 177, 71.

Langmuir 2010, 26(12), 10115–10121

Zhu and Huang

Article

Figure 6. 51V NMR spectra of octylamine (C8)-intercalated R-VOPO4 3 2H2O: (A) MAS at 14.1 T, (B) expanded MAS spectra to show isotropic shifts, and (C) static echos at 9.4 and 14.1 T. Figure 5.

51

V static echo (A) and MAS (B) NMR spectra of hexylamine (C6)-intercalated R-VOPO4 3 2H2O at fields of 9.4 (left column) and 21.1 T (right column). (C) The expanded MAS spectra show isotropic shifts.

present study, the intercalation was carried out at room temperature without heating by using hydrated R-VP, a very mild condition that is similar to that employed by Beneke and Lagaly.18 To understand the intercalation better, simple calculations were carried out on a small octahedral cluster, (H2O)V(dO)(OH)4. The results (Supporting Information Table S1) indicate that replacing the coordinated water molecule with NH3 results in a significant increase in the CQ value but very little change in Ω. However, small alternations in V-O distances in VdO and VOH2 groups lead to a significant change in Ω. On the basis of the intercalation conditions employed in this work and the calculations on the simplified model, we tentatively suggest that amine insertion might occur via hydrogen bonding of the amine headgroup with the VdO group and the water molecule coordinated to the V center. However, the possibility of the direct coordination of amine to the V center cannot be ruled out completely at this point, and further investigation into the mechanism of interaction between the V center and the amine molecule is needed. The 51V MAS and static spectra (Figure 5) of hexylamineintercalated R-VP (C6-VP) were similar to those of C5-VP. The following tensor parameters were extracted from the spectra: δiso=-550 ppm, Ω=488 ppm, κ=0.12, CQ=2.7 MHz, ηQ=0.80, and Euler angles (R, β, γ)=(30, 140, 0). The similarity in the EFG and the CS tensor parameters (Table 1) between C5-VP and C6-VP indicates that the addition of a methylene group has little effect on the V environment. The XRD pattern shows that in addition to highly crystalline C6-VP there is a small amount of layered material existing as an impurity. The origin of this second phase is not clear. The line shape of the 51V central transition (Figure 6) in the spectra of octylamine-intercalated R-VP (C8-VP) looks distinctly different from those of C5-VP and C6-VP. The MAS spectra reveal that there are two 51V species with isotropic chemical shifts at -551 ppm (δiso1) and -540 ppm (δiso2). The NMR parameters of the resonance with δiso1 at -551 are very similar to those of C6-VP (Table 1). The second signal at -540 ppm has the following tensor parameters: δiso =-540 ppm, Ω=1006 ppm, κ= 0.93, CQ=2.7 MHz, ηQ=0.80, and Euler angles (R, β, γ)=(0, 50, 90). The corresponding XRD pattern (Figure 2) clearly shows Langmuir 2010, 26(12), 10115–10121

that the sample is a mixture of two layered phases with different basal spacings. Apparently, the two 51V signals originate from these two phases. The exact assignments of these two resonances will be discussed shortly. For the R-VPs intercalated with longer alkylamines including dodecylamine (C12) and hexadecylamine (C16), the line shapes of the static spectra (Figure S3) are similar to that of C8-VP. The static spectra of both intercalates can be fitted reasonably well with two resonance signals using the same sets of parameters for two signals in C8-VP but with different intensity ratios. The similarity between the C12-VP and C16-VP spectra suggests that the inclusion of the two amines has a similar effect on the V local structure. Previous work has shown that the basal spacing of n-alkylamine-intercalated R-VP can be linearly correlated with the chain length of the amines (i.e., the number of carbon atoms (NC) in the alkyl chains).18-21 The dependence of the basal spacing of intercalated VPs derived from the (001) reflection in the XRD patterns on NC is illustrated in Figure 7A. Fitting the four data points of C3-VP, C5-VP, C6-VP, and the C8-VP phase with a smaller d spacing in the mixture yields a good straight line. The slope of the line is 1.78, meaning that the addition of each methylene (-CH2-) group results in an increase in the d spacing by 1.78 A˚. The three points of C16-VP, C12-VP, and the second phase of C8-VP with a larger basal spacing also fall on a straight line with a larger slope of 2.58. Early work suggested that when long-chain amines and alkanols are intercalated into R-VP the guest molecules could, in principle, adopt either monomolecular or bimolecular layer arrangements.19-21 For a monomolecular arrangement, if the alkyl chain of an amine molecule is in a fully stretched all-trans conformation and orients itself perpendicular to the VP basal plane, then theoretically the addition of each methylene group should lead to an increase in the basal spacing by 1.28 A˚, whereas the theoretical increase in the d spacing with the addition of each CH2 group in a bimolecular layer model is 2.56 A˚. The fact that the slopes of both straight lines are greater than 1.28 clearly implies that in the present work, regardless of the chain length, all of the amines form bimolecular layers, which is consistent with the finding reported in the literature by others.19-21 For the amines with NC g 8, the slope of 2.58 is very close to the theoretical value of 2.56, indicating that for these amines their alkyl chains adopt the all-trans configuration and are perpendicular to the VP basal plane. For the amines with NC e 8, the slope (1.78) is smaller than 2.58, suggesting that the amine molecules DOI: 10.1021/la100232s

10119

Article

Figure 7. Plots of (A) basal spacing and (B) 51V CSA span (Ω) against the number of carbons (NC) in the amines with a tilted orientation.

forming the bimolecular layer have their molecular axes tilted with respect to the VP layer, resulting in a smaller slope. Furthermore, following the approach of Bene et al.,22 the angle formed between the molecular axis (along the C-C chain direction) and the VP plane is estimated to be 44. Another factor that may also cause the slope to be smaller than 2.58 is the conformation of the alkyl chain. As shown previously by 13C NMR and IR studies, the amine chain can also adopt the disordered gauche conformation, which is particularly true for the amine with shorter chains.21 Realistically, for the amines with NC e 8, it is not unreasonable to assume that the alkyl chains are both tilted and conformationally disordered. The packing of the amines with NC g 8 between the VP layers is more ordered. As mentioned earlier, the XRD pattern of C8-VP shows two strong low-angle peaks at 4.02 and 4.79 (Figure 2E) corresponding to the two intercalated VP phases with different basal spacings of 26.0 and 21.5 A˚, respectively. The intercalated VP with a d spacing of 26.0 A˚ can be assigned to C8-VP with the octylamine perpendicular to the basal plane whereas the phase with a basal spacing of 21.5 A˚ can be assigned to the C8-VP phase where the octylamine molecules have a tilted orientation. It seems that octylamine is the transition point from the tilted to the perpendicular configuration. On the basis of the above discussion and the NMR parameters in Table 1, the two 51V signals in the spectra of C8-VP can now be assigned: the signal at -551 ppm is apparently due to the V centers interacting with the octylamine molecules taking tilted orientations in the C8-VP phase with a smaller basal spacing. The signal at -540 ppm originates from the V atoms associated with the octylamine perpendicular to the basal plane with a larger d spacing. As mentioned earlier, 51V NMR spectra of C12-VP and 10120 DOI: 10.1021/la100232s

Zhu and Huang

C16-VP also contain the same two signals at -551 and -540 ppm as observed in C8-VP. However, unlike C8-VP, the XRD patterns show that both C12-VP and C16-VP exist as a single phase with high crystallinity. In each case, the first three strong low-angle reflections can be indexed as (001), (002), and (003). Thus, it seems that for C12-VP and C16-VP two different V environments exist in the same basal plane: the signal at -540 ppm represents the majority of the V centers bound to the amine molecules with the all-trans configuration and oriented perpendicularly to the inorganic layer. These amine molecules determine the basal spacing of the intercalated phase. Within the basal plane, there also exists another type of V atom (about 33%) associated with the amine molecules adopting the tilted orientation, and the existence of this type of V environment is not apparent in the XRD pattern. This is because XRD is sensitive to long-range ordering whereas NMR is more sensitive to the local environment. As a result, the existence of two V domains within the same layer can be detected by 51V NMR. Figure 7B shows a plot of 51V spans (Ω) as a function of NC for the amines with a tilted orientation. The span initially decreases with increasing NC, but after the number of carbon atoms in the chain reaches six, the span values remain essentially constant and further increases in NC do not induce a significant change in the span. This is understandable considering that the CS interaction is sensitive to short-range interaction and is mostly influenced by the local arrangements of electrons. Therefore, it is not expected to be strongly affected by the long-range effect. In the present case, hydrogen bonding with the amine initially has a large effect on the V bonding environment and produces a larger change in the CSA. When the alkyl chains are short, adding each -CH2- group will have a noticeable influence on the hydrogen bonding and therefore on the V-O bond strengths and — O-V-O angles. However, when the chains reach a certain length, the addition of another methylene group will not significantly affect the V bonding environment. Table 1 shows that the dependence of 51V skew (κ) as a function of NC for the amines with a tilted orientation is similar to that of the span. The CS tensor symmetry of the parent materials is almost axially symmetric and decreases with an increasing number of carbon atoms in the chain. The skew value is 0.12 for amines with 5-16 carbons. This observation suggests that the local symmetry around the V center decreases upon intercalation with amine. The changes in — O-V-O angles and V-O bond lengths upon intercalation result in the reduction of the CS tensor symmetry. The 31P MAS spectra of R-VP and amine-intercalated phases are shown in Figure S4. All of the spectra exhibit a single P peak (except for the two resonances that were observed in C5-VP). The 31 P isotropic chemical shift gradually moved from 9.5 in R-VP to 3.3 ppm in C8-VP upon intercalation, and it does not change with further increases in the chain length. Although small, the shifts in the 31P signals reflect the interaction of the amines with the VO6 octahedra. The fact that C8-VP, C12-VP, and C16-VP all have the same 31P shift suggests that the conformation and orientation of the long-chain amine molecules have little effect on the P environment. This observation is different from the situations for alkylamine-intercalated layered R-Zr(HPO4)2H2O (R-ZrP) where intercalation usually results in a dramatic change in the spectrum with multiple P sites and a much larger dispersion of the chemical shift. This is due to the fact that for R-ZrP the intercalation occurs via the formation of hydrogen bonds between amine and P-OH groups from the inorganic layer.48 In the present case, the intercalation of R-VOPO4 3 2H2O involves hydrogen (48) Maclachlan, D. J.; Morgan, K. R. J. Phys. Chem. 1992, 96, 3458.

Langmuir 2010, 26(12), 10115–10121

Zhu and Huang

bonding with axial V-O groups, leaving the PO4 tetrahedron relatively unaffected. Therefore, for the intercalated VPs, 51V NMR seems to be a more direct probe than 31P NMR. In summary, R-VP and its alkylamine-intercalated derivatives were investigated using powder XRD and, in particular, solidstate 51V NMR. The XRD results show that the amines form bimolecular layers upon intercalation. For the amines with NC < 8, the alkyl chains are tilted and perhaps have a disordered gauche conformation. For the amines with NC > 8, a majority of the amines adopt an orientation where the C-C chains are perpendicular to the basal planes. However, the tilted orientation also exists. For NC = 8, two intercalated phases coexist: the intercalated phase with the larger basal spacing has the amine molecules taking the perpendicular orientation and the phase with the smaller d spacing contains the amine molecules with a tilted configuration. At this point, it is not clear why the intercalation of R-VP with octylamine results in phase separation. The 51 V CS parameters (δiso, Ω, and κ) are sensitive to the intercalation as well as the orientation of the guest molecules in the host. The EFG tensor, however, is not strongly influenced by the intercalation with amines.

Langmuir 2010, 26(12), 10115–10121

Article

Acknowledgment. Y.H. acknowledges financial support from the NSERC of Canada in the form of a research grant and the CFI for equipment grants. Funding from the Canada Research Chair program is also gratefully acknowledged. Access to the 900 MHz NMR spectrometer was provided by the National Ultrahigh Field NMR Facility for Solids (Ottawa, Canada), a national research facility funded by the Canada Foundation for Innovation, the Ontario Innovation Trust, Recherche Quebec, the National Research Council Canada, and Bruker BioSpin and managed by the University of Ottawa (http://www.nmr900.ca). We thank Drs. Terskikh, Pawsey, and Kirby for technical assistance with the NMR experiments and Prof. Wasylishen for the WSOLIDS1 software. We thank five anonymous reviewers for their insightful comments and helpful suggestions. Supporting Information Available: 51V NMR spectra of V2O5; 51V and 31P NMR spectra of several amine-intercalated R-VPs; and the results of theoretical calculations on a model cluster. This material is available free of charge via the Internet at http://pubs.acs.org.

DOI: 10.1021/la100232s

10121