ohio2/yjy-yjy/yjy-yjy/yjy99907/yjy9002d07z xppws 23:ver.3 4/30/09 14:51 Msc: jp-2009-03500q TEID: pjp00 BATID: jp6d233

J. Phys. Chem. C XXXX, xxx, 000 A

1 Optimized Synthesis and Structural Characterization of the Borosilicate MCM-70

† ,† † #,‡ ‡ 2 Dan Xie, Lynne B. McCusker,* Christian Baerlocher, Lisa Gibson, Allen Burton, and § 3 Son-Jong Hwang 4 Laboratory of Crystallography, ETH Zurich, CH-8093 Zurich, Switzerland, CheVron Energy Technology Co., 5 Richmond, California 94802, The DiVision of Chemistry and Chemical Engineering, California Institute of 6 Technology, Pasadena, California 91125, USA

7 ReceiVed: April 16, 2009; ReVised Manuscript ReceiVed: April 24, 2009

8 A structure analysis of the borosilicate zeolite MCM-70, whose synthesis had been patented in 2003, was 9 reported in 2005. Unfortunately, that structure analysis was somewhat ambiguous. Anisotropic line broadening 10 made it difficult to model the peak shape. Some peaks in the electron density map could not be interpreted 11 satisfactorily, the framework geometry was distorted, and MAS NMR results were partially contradictory. In 12 an attempt to resolve some of these points, an optimization of the synthesis was undertaken, and the structure 13 was reinvestigated. The structure was solved from synchrotron powder diffraction data collected on an as- 14 synthesized sample (Pmn21, a ) 13.3167(1) Å, b ) 4.6604(1) Å, c ) 8.7000(1) Å) using a powder charge- 15 flipping algorithm. The framework topology, with a 1-dimensional, 10-ring channel system, is identical to 16 the one previously reported. However, the B in this new sample was found to be ordered in the framework, + 17 fully occupying one of the four tetrahedral sites. Two extra-framework K ion positions, each coordinated to 18 five framework O atoms and one water molecule, were also found. The solid state 29Si, 11B and 1H NMR 19 results are fully consistent with this ordered structure.

20 Introduction deionized water. Next, 3.50 g of Ludox-30 colloidal silica 47 was mixed with the solution to create a uniform gel. The 48 21 In 2005, Dorset and Kennedy reported the crystal structure 22 of the borosilicate zeolite MCM-70,1 whose synthesis had been linear was then capped and placed within a Parr Steel 49 23 patented 2 years earlier.2 However, the structure analysis was autoclave reactor. The autoclave was fixed in a rotating spit 50 24 complicated by the fact that the peak shapes in the powder within an oven heated to 160 °C for 11 days. The solid 51 25 diffraction pattern exhibited anisotropic line broadening and products were then recovered from the cooled reactor by 52 26 were difficult to model. During the course of the Rietveld re- vacuum filtration and washed with at least 250 mL of 53 27 finement, extra-framework species, whose positions were not deionized water. The X-ray powder diffraction pattern was 54 28 easy to interpret, were found in the electron density maps gen- consistent with the data provided in the patent. 55 29 erated for both as-synthesized and dehydrated samples. The Chemical Analysis. Chemical analysis was performed by 56 30 geometries of the framework structures in both cases were Galbraith Laboratories, Inc. using inductively coupled plasma 57 31 distorted. There also appeared to be a large amount of K in the (ICP) methods. 58 32 structure for the relatively small amount of B in the framework. NMR. Solid state 1H, 11B, and 29Si MAS and CPMAS NMR 59 33 Finally, although the 29Si MAS NMR spectrum was indicative spectra were recorded using a Bruker DSX-500 spectrometer 60 34 of a nonrandom distribution of B in the framework structure, (11.7 T) and a boron-free Bruker 4 mm CPMAS probe. The 61 35 this was not apparent in the crystal structure analysis. With the 36 aim of clarifying some of these details, an optimization of the spectral frequencies were 500.23 MHz, 160.50 MHz, and 99.4 62 1 11 29 37 synthesis of MCM-70 was undertaken. The improved sample MHz for H, B, and Si nuclei, respectively. NMR shifts (in 63 1 29 38 was then examined using MAS NMR and synchrotron powder parts per million, ppm) for H and Si nuclei were externally 64 11 39 diffraction techniques. referenced to tetramethylsilane (TMS), and those for B, to 65 BF3 · O(CH2CH3)2. Powder samples were packed in a ZrO2 rotor 66 40 Experimental Section under ambient conditions and spun at 2-15 kHz. A typical 11B 67 68 41 Synthesis. The MCM-70 sample was synthesized according MAS NMR spectrum was obtained after a 0.3 µs single pulse < 1 42 to a procedure given in the original patent2 with only minor ( π/12) with the application of a strong H decoupling pulse 69 329 43 modification. In a 23-mL Teflon linear, 0.22 g of boric acid, of the two-pulse phase modulation (TPPM) scheme. Si Bloch 70 44 0.44 g of solid potassium hydroxide (88%), and 1.08 g of decay spectra were obtained using a 4 µs-π/2 pulse and a recycle 71 29 1 45 N,N′-diisopropyl-N,N′-dipropylbicyclo[2.2.2]oct-7-ene-2,3/ delay time of 2000 s at spinning speeds of 8 kHz. Si{ H} and 72 46 5,6-dipyrrolidinium diiodide were dissolved in 7.0 g of 11B{1H} CPMAS spectra were collected at variable contact times 73 from 0.2 m to 5 ms. 74 * To whom correspondence should be addressed. Phone: +41-44-632- Powder Diffraction Data Collection. High-resolution X-ray 75 3721. Fax: +41-44-632-1133. E-mail: [email protected]. † ETH Zurich. powder diffraction data were collected on the Swiss-Norwegian 76 F1 ‡ Chevron. Beamline (SNBL) at the European Synchrotron Radiation 77 § California Institute of Technology. # Present address: Cambridge University Chemical Laboratory, Lensfield Facility (ESRF) in Grenoble, France. Data collection parameters 78 Road, Cambridge, CB2 1EW, U.K. are given in Table 1. 79 10.1021/jp903500q CCC: $40.75  XXXX American Chemical Society

PAGE EST: 5.9 ohio2/yjy-yjy/yjy-yjy/yjy99907/yjy9002d07z xppws 23:ver.3 4/30/09 14:51 Msc: jp-2009-03500q TEID: pjp00 BATID: jp6d233

B J. Phys. Chem. C, Vol. xxx, No. xx, XXXX Xie et al.

TABLE 1: X-ray Powder Diffraction Data Collection MAS NMR spectrum (delay time of 2000 s) at -95.8, -98.5, 91 - synchrotron facility SNBL (station B) at ESRF and 103.0 ppm with the relative intensities of 1:2:2 is 92 consistent with the results reported by Dorset and Kennedy.1 93 wavelength 0.497 19 Å - - The small peak at 107.7 ppm is attributed to a minor quartz 94 diffraction geometry Debye Scherrer ∼ analyzer crystal Si 111 impurity ( 2%). The spectral resolution for both MAS and 95 F2 sample rotating 1.0 mm capillary CPMAS NMR of our material appear to be superior to that in 96 2θ range 1-43° the previous report, where a broadening of the MAS spectrum 97 step size 0.003° 2θ was observed and attributed to local disorder in the material or 98 time per step sample heterogeneity. It was recognized by Dorset and Kennedy, 99 1.0-10.5°2θ 1.5 s however, that the 29Si NMR spectra are indicative of a 100 10.5-22.5°2θ 3.0 s 22.5-43.0°2θ 4.5 s nonrandom distribution of B in the framework structure, with 101 boron adopting a specific T-site rather than being distributed 102 uniformly over all T-sites. 103 80 Results 1H static and MAS NMR spectra are shown in Figure 1b and 104 81 Chemical Analysis. The initial ICP analysis of the MCM- are informative in identifying the protons in the sample. The 105 1 82 70 sample yielded 34.4 wt % Si, 0.46 wt % B, and 1.77wt % presence of two kinds of protons is clearly visible in the H 106 83 K, giving a Si/B ratio of 28.5 and a Si/K ratio of 27.2. A repeat MAS spectrum at higher spinning rate (13 kHz). A broad peak 107 84 measurement of the same sample resulted in 39.6 wt % Si, 2.3 at 4.1 ppm is attributed to surface water, and its signal intensity 108 85 wt % B, and 12.1wt%KorSi/B 6.6 and Si/K 4.6. The vast is ∼20% of the total hydrogen content. The surface water could 109 86 differences between these two analyses indicate that large errors be removed readily by dehydration at near 100 °C. The 110 87 are involved, so they were not used further. remaining 80% of the major hydrogen species is represented 111 88 NMR. Figure 1a shows the 29Si MAS and CPMAS spectra as a peak at 2.96 ppm, with numerous spinning sidebands and 112 89 of an MCM-70 sample that was calcined and then stored at a relatively sharp line width (∼300 Hz). It is also remarkable 113 90 ambient conditions. The appearance of three sharp peaks in the that no peaks near ∼1.7-2.0 ppm due to surface silanol groups 114

Figure 1. Multinuclear MAS NMR spectra of calcined MCM-70. (a) 29Si MAS and CPMAS NMR (using cross-polarization contact time of 1 and 10 ms); (b) 1H static and MAS at two different spinning rates (2.5 and 13 kHz); (c) 11B MAS NMR and CPMAS NMR (contact time ) 1.0 ms) of the center band; (d) Experimental and fitted 11B MAS NMR spectra showing spinning sidebands arising from all transitions of the quadrupole nucleus (I ) 3/2). ohio2/yjy-yjy/yjy-yjy/yjy99907/yjy9002d07z xppws 23:ver.3 4/30/09 14:51 Msc: jp-2009-03500q TEID: pjp00 BATID: jp6d233

Synthesis and Structure of the Borosilicate MCM-70 J. Phys. Chem. C, Vol. xxx, No. xx, XXXX C

Figure 2. Powder charge-flipping electron density map for as-synthesized MCM-70 generated with Pmn21 symmetry imposed. The stick model of the final structure is overlaid for comparison (Si, blue; B, light blue; O, red; K, green).

TABLE 2: Crystallographic Data for MCM-70 TABLE 3: Selected Bond Distances (Å) and Angles (°) for MCM-70 chemical composition |K2(H2O)2|[Si10B2O24] Si-O, min 1.58 B-O, min 1.47 unit cell max 1.66 max 1.52 a 13.3167(1) Å avg 1.61 avg 1.49 b 4.6604(1) Å O-Si-O, min 104.0 O-B-O, min 108.3 c 8.7000(1) Å max 114.3 max 110.1 space group Pmn2 1 avg 109.5 avg 109.5 no. observations 4707 Si-O-Si, min 133.7 Si-O-B, min 126.8 no. contributing reflections 677 max 158.0 max 133.0 no. geometric restraints 41 avg 142.0 avg 130.8 Si-O 1.60(1) Å 11 K(1)-O(4) ×2 2.99(1) K(2)-O(1) ×2 3.07(1) B-O 1.48(1) Å 3 K(1)-O(6) ×2 3.08(1) K(2)-O(2) ×2 3.04(1) O-Si-O 109.5(1.0)° 16 K(1)-O(7) 2.86(2) K(2)-O(3) 3.00(2) O-B-O 109.5(1.0)° 4 K(1)-Ow(1) 2.77(2) K(2)-Ow(2) 2.77(2) Si-O-Si 145(8)° 5 Si-O-B 145(8)° 2 1 no. structural parameters 37 hydrogen bonding is negligible. The characteristic H NMR line 135 no. profile parameters 8 shape did not change when the surface water was removed by 136 1 RF 0.072 evacuation at 120 °C, so the major H MAS NMR signal at 137 Rwp 0.135 2.96 ppm can probably be attributed to well-separated ordered 138 R 0.047 exp water molecules. Dorset and Kennedy assigned a similar peak 139 at 3.15 ppm to B-OH groups, but no powder pattern of dipole 140 115 (Si-OH) are observed.4 When the spinning rate was reduced coupling was shown in that paper. 141 116 to 2.5 kHz, further broadening of the spinning sidebands did With the information from 1H NMR, the 29Si CPMAS NMR 142 117 not occur, while the well-known powder pattern of Pake spectra can be analyzed further. The nonrandom ordering 143 118 doublets5,6 was unveiled. The static spectrum (0 kHz) also behavior can also be seen in these spectra from the nonuniform 144 119 confirms the line shape. The characteristic spectrum is a typical signal growth of the three peaks for two different cross- 145 120 sign of the presence of isolated and dipolar coupled spin pairs polarization contact times (1 and 10 ms). We believe that the 146 T2 121 in powders. The most plausible case in the MCM-70 sample is proximity of protons to T-sites is unveiled briefly in this simple 147 1 122 either two H nuclei in H2O or two OH groups in close proximity test. The T-site responsible for the middle peak at -98.5 ppm 148 123 to one another. The distance between the 1H spin pair was seems to be farther away from the crystalline water than the 149 124 measured to be 1.65 Å from the strength of the 1H dipole other two. 150 H-H 11 125 coupling: ΗD ) 26.7 kHz (∼40 kHz separation of the Pake Figure 1c and d show B MAS spectra for the center band 151 126 doublet). The presence of two isolated hydroxyl groups (Si-OH and spinning sidebands from all transitions of the quadrupole 152 127 or B-OH) at such a close distance is unlikely, but isolated water nuclei (I ) 3/2), respectively. The single peak at -1.47 ppm 153 128 molecules could explain the 1H NMR spectra. It is also possible with a line width of 80 Hz in the 11B MAS NMR spectrum is 154 129 to conclude that the water molecules must be part of a highly associated with tetrahedral boron in the framework, as pointed 155 130 ordered arrangement with the oxygen atom associated with a out by Dorset and Kennedy. The same peak observed via 11B 156 131 tight coordination, resulting in a slight increase in the O-H CPMAS NMR indicates that the boron site is near the ordered 157 132 bond distance compared to the normal one. The mobility of the water. The spinning sidebands were fitted with Qusar software7 158 T3 133 water molecule is certainly limited. Although surface water is to extract the quadrupole parameters: Qcc ) 182 kHz, ηQ ) 159 134 present (see above), its interaction with the ordered water via 0.45, where Qcc and ηQ are the quadrupole coupling constant 160 ohio2/yjy-yjy/yjy-yjy/yjy99907/yjy9002d07z xppws 23:ver.3 4/30/09 14:51 Msc: jp-2009-03500q TEID: pjp00 BATID: jp6d233

D J. Phys. Chem. C, Vol. xxx, No. xx, XXXX Xie et al.

Figure 3. Observed (top), calculated (middle), and difference (bottom) profiles for the Rietveld refinement of as-synthesized MCM-70. The profiles in the inset have been scaled up by a factor of 5 to show more detail. Tick marks indicate the positions of reflections.

161 and the asymmetry parameter, respectively.7 All 11B MAS NMR 162 results again indicate that the tetrahedral boron is in a well- 163 defined crystalline framework and located at a specific site. 164 Structure solution. The diffraction pattern could be indexed 165 on a primitive orthorhombic unit cell (a ) 13.316 Å, b ) 4.658 166 Å, c ) 8.698 Å) using the program TREOR8 implemented in 167 the software CMPR.9 A careful examination of the diffraction 168 pattern indicated that h0l reflections with h + l ) 2n + 1 were 169 systematically absent, so as in the original work of Dorset and 170 Kennedy,1 the two most probable space groups were expected 171 to be Pmnm (centrosymmetric, standard setting Pmmn)orPmn21 172 (noncentrosymmetric). 173 Assuming the centrosymmetric space group Pmnm, reflection 174 intensities were extracted from the powder pattern to a minimum 175 d spacing of 0.77 Å (ca. 37.6° 2θ) using the program 176 EXTRACT10 in the XRS-82 suite of programs.11 To optimize 177 the partitioning of overlapping reflections as far as possible, Figure 4. The framework structure of MCM-70 showing the edge- 178 the extracted intensities were repartitioned using the fast iterative sharing 4-ring pairs and the 10-ring channels. Bridging O atoms have 179 Patterson squaring (FIPS) method.12 These intensities were then been omitted for clarity. The B position is shown in gray. 180 expanded to the space group P1 and used as input to the powder 181 charge-flipping (pCF) structure solution algorithm13 that is maps generated in P1 appeared to be quite interpretable, those 199 182 implemented in the program Superflip.14 In this implementation, obtained from the second series of pCF runs with symmetry 200 183 the intensity repartitioning procedure for overlapping reflections averaging were slightly better. The 10 best maps were averaged, 201 F4 184 is coupled to a second modification of the electron density map and from this map, not only could the positions of all atoms in 202 185 based on histogram matching.15 The reference histogram used the asymmetric unit (4 T atoms, 2 extra-framework atoms, and 203 186 for this step simply reflects the chemical composition of the 7 O atoms) be determined directly from the 15 highest peaks 204 187 material. in the density map, but also the peak heights were consistent 205 188 Initially, no symmetry was imposed during the pCF runs with the different types of atoms (Figure 2). The electron density 206 189 because the space group was not clear. Instead, the method at one of the tetrahedrally coordinated framework atom (T-atom) 207 190 proposed by Palatinus and van der Lee16 for determining the sites was significantly lower than at the other three, so it was 208 191 space group symmetry from the final electron density map assumed that this site probably contained a substantial amount 209 192 generated in P1 was applied. In 100 pCF runs, the space group of B. The T-O bond distances from this site were also found 210 193 Pmn21 was found 83 times, but Pmnm was not found at all. to be shorter than those from the other three, as would be 211 194 Furthermore, 8 of the 10 best density maps (according to the expected for a higher B content (B-O, ca. 1.47 Å; Si-O, ca. 212 17 195 Superflip R value) showed Pmn21 symmetry. This is a strong 1.61 Å). 213 196 indication that the noncentrosymmetric space group Pmn21 is Structure Refinement. To begin with, the four T atoms were 214 197 the correct one. Another 100 pCF runs were then performed refined as Si atoms with occupancies set arbitrarily at 0.95, 0.95, 215 198 with Pmn21 symmetry imposed. Although the electron density 0.95, and 0.3 to reflect the probable high B content in the latter. 216 ohio2/yjy-yjy/yjy-yjy/yjy99907/yjy9002d07z xppws 23:ver.3 4/30/09 14:51 Msc: jp-2009-03500q TEID: pjp00 BATID: jp6d233

Synthesis and Structure of the Borosilicate MCM-70 J. Phys. Chem. C, Vol. xxx, No. xx, XXXX E

Figure 5. The structure of MCM-70 viewed down the b-axis showing a possible arrangement of the extra-framework atoms. Three channels are shown with K+ ions at K(1) and water molecules at Ow(1), whereas the other two are shown with K+ ions at K(2) and water molecules at Ow(2) to reflect the 57% and 43% occupancy factors.

occupied with B and that the other three were pure Si. The total 232 occupancy of the K(1) and K(2) positions, however, consistently 233 refined to a value above 1.0. Reasoning that the excess electron 234 density might be from water (a K-O distance of 2.75 Å would 235 be reasonable), water was added at each K position. Because a 236 previous difference electron density map with the K atoms 237 removed from the model had shown only single spherical peaks 238 at the two positions, the water molecules were constrained to 239 have coordinates identical to those of the respective K atoms, 240 and the total occupancy at each site (K(1)/Ow(2) and K(2)/ 241 Ow(1)) was constrained to be not more than 1.0. This refinement 242 not only reduced the R values significantly but also yielded a 243 chemically reasonable structure. The occupancies at K(1) and 244 K(2) refined to 0.58 and 0.42, respectively. The remaining 245 electron density at each position refined to 0.42 and 0.58 for 246 Ow(2) and Ow(1), respectively. 247 Refinement of this structural model with the chemical 248 formula |K2(H2O)2|[Si10B2O24] converged with RF ) 0.072 249 and Rwp ) 0.135. All displacement factors were refined 250 + Figure 6. The coordination spheres of the K ions at K(1) and K(2). isotropically, and those for similar atoms were constrained 251 to be equal to keep the number of parameters to a minimum. 252 Neutral scattering factors were used for all atoms. Details 253 of the refinement are given in Table 2, and selected bond 254 217 The distance between the two extra-framework-atom positions distances and angles, in Table 3. A cif file with the final 255 218 (ca. 2.75 Å) was too short to allow simultaneous occupation atomic parameters is provided in the Supporting Information. 256 F6 219 by K, so each of these sites was initially assigned as K with an The fit of the profile calculated from the final model to the 257 220 occupancy of 0.5. This model appeared to be consistent with experimental data is shown in Figure 3. 258 221 the NMR data and to give a reasonable fit to the diffraction 222 pattern (RF ) 0.103, Rwp ) 0.310, Rexp ) 0.048), so it was used 223 as a starting point for Rietveld refinement using the XRS-82 Discussion 259 11 224 suite of programs. Geometric restraints were placed on the Structure. The framework structure of MCM-70 can be 260 225 bond distances and angles of the framework atoms throughout described in terms of pairs of edge-sharing 4-rings. These pairs, 261 226 the refinement (see Table 2), but their relative weight with arranged in a centered (in projection) rectangle in the ac plane, 262 227 respect to the powder diffraction data was reduced as the are linked together to form 10-rings, and are connected to one 263 228 refinement progressed. another along the b axis in a zigzag manner to form a three- 264 229 Once the positions had been refined, the occupancies of the dimensional, four-connected net with a one-dimensional, 10- 265 + 230 Si and K atom positions were also allowed to vary. These ring channel system (Figure 4). The extra-framework K ions 266 + 231 refinements indicated that the fourth T site was, indeed, fully are located in the 10-ring channels. Because the K ions at K(1) 267 PAGE EST: 5.9 ohio2/yjy-yjy/yjy-yjy/yjy99907/yjy9002d07z xppws 23:ver.3 4/30/09 14:51 Msc: jp-2009-03500q TEID: pjp00 BATID: jp6d233

F J. Phys. Chem. C, Vol. xxx, No. xx, XXXX Xie et al.

268 and K(2) cannot coexist in the same channel, 58% of the Conclusions 329 + 269 channels have K ions at K(1) and water molecules at Ow(1), + The framework structure for the borosilicate zeolite MCM- 330 270 and 42% have K ions at K(2) and water molecules at Ow(2) 70 originally proposed by Dorset and Kennedy, with a 1-di- 331 271 (Figure 5). The two K sites are very similar to one another + mensional, 10-ring channel structure, has been confirmed. 332 272 (related by a pseudo inversion center). In both cases, the K Optimization of the synthesis conditions for this material resulted 333 273 ions are located opposite 8-ring pockets in the walls of the 10- in a more crystalline/homogeneous sample, and structural 334 274 ring channels, where they coordinate to five O atoms of the characterization of this sample using X-ray powder diffraction 335 275 framework and one water molecule (Table 3, Figure 6). The and solid state NMR techniques has allowed some of the 336 276 water molecules, in turn, are at hydrogen bonding distances from ambiguities in the earlier study to be resolved. In particular, 337 277 framework O atoms. the boron has been shown to be completely ordered, fully 338 278 The structure is very similar to that reported earlier by occupying one of the four tetrahedral sites in the framework 339 279 Dorset and Kennedy.1 The major differences are that (1) the structure. The remaining three sites are pure Si. For each B atom 340 280 B atoms in the framework structure of the material used in + + in the framework, there is a K counterion in the channels. These 341 281 this study are clearly ordered and (2) the K ions are found + K ions partially occupy two very similar sites in different 10- 342 282 in two (rather than one) closely related, but crystallographic ring channels, where they bond to five framework O atoms and 343 283 distinct, positions in the 10-ring channels. The sample one water molecule. The 29Si, 11B, and 1H solid state NMR 344 284 described here has a much higher percentage of B in the results confirm both the ordering of the B atoms in the 345 285 framework structure (Si/B 5:1 vs 40:1 in the previous work). framework and the ordered arrangement of isolated water 346 286 Apparently, this higher B concentration has led to a more molecules in the channels. 347 287 crystalline material with a simple stoichiometry 288 (|K2(H2O)2|[Si10B2O24]), a clear ordering of B in the frame- + Acknowledgment. We thank the beamline scientists at the 348 289 work structure, and the appropriate number of K ions in SNBL at the ESRF in Grenoble, France, for their assistance 349 290 the channels to balance the charge of the borosilicate with the powder diffraction measurements. The NMR facility 350 291 framework. at Caltech was supported by the National Science Foundation 351 292 The framework topology itself is centrosymmetric, but (NSF) under Grant No. 9724240 and supported in part by the 352 293 because the B atoms are ordered, the inversion center is violated MRSEC Program of the NSF under Award No. DMR-0520565. 353 294 and the structure crystallizes in the noncentrosymmetric space This work was funded in part by the Swiss National Science 354 295 group Pmn21. It is interesting that although there was no Foundation. 355 296 crystallographic evidence for an ordering of the B atoms in the 297 Dorset and Kennedy study, their structure solution using the Supporting Information Available: Crystallographic in- 356 298 program FOCUS18 was also found with this space group rather formation file for MCM-70 (CIF). This material is available 357 299 than the centrosymmetric Pmnm. free of charge via the Internet at http://pubs.acs.org. 358 300 MAS NMR. Assignment of the three Si sites in the structure 301 (see Figure 4) to the 29Si MAS NMR signals appears to be 302 straightforward. The signal at -103.0 ppm can be attributed to References and Notes 359 - - - 303 Si(1), with no O B bonds; that at 98.5 ppm to Si(2), with (1) Dorset, D. L.; Kennedy, G. J. J. Phys. Chem. B 2005, 109, 13891. 360 304 one -O-B bond; and that at -95.8 ppm to Si(3), with two (2) Dinghra, S. S.; Weston, S. C. US Patent 6,656,268, 2003. 361 305 -O-B bonds. The multiplicities of these three sites (4, 4, and (3) Bennett, A. E.; Rienstra, C. M.; Auger, M.; Lakshmi, K. V.; Griffin, 362 306 2, respectively), correspond well to the observed intensities. R. G. J. Chem. Phys. 1995, 103, 6951. 363 (4) Pfeifer, H.; Ernst, H. In Annual Reports on NMR Spectroscopy; 364 307 29 From the Si NMR spectrum, the Si/B ratio would be calculated Academic Press: London, 1994, Vol. 28, p. 91. 365 308 to be ([Si(1)] + [Si(2)] + [Si3)])/([Si(2)]/4 + [Si(3)]/2) ) 5/(2/4 (5) Pake, G. E. J. Chem. Phys. 1948, 16, 327. 366 309 + 1/2) ) 5, which is completely consistent with the structure (6) Abragam, A. The Principles of Nuclear Magnetism; Clarendon 367 310 described above. The distances between the water molecules at Press: Oxford, 1961, Chapter 4. 368 (7) Amoureux, J. P.; Fernandez, C.; Carpentier, L.; Cochon, E. Phys. 369 311 Ow(1), the more highly occupied water site, and the three Si 29 Status Solidi A 1992, 132, 461. 370 312 sites are also consistent with the Si CPMAS NMR spectrum. (8) Werner, P. E.; Eriksson, L.; Westdahl, M. J. Appl. Crystallogr. 371 313 The Si atoms at Si(2) (signal at -98.5 ppm) are the farthest 1985, 18, 367. 372 314 away from Ow(1), and those at Si(3) (signal at -95.8 ppm) are (9) Toby, B. H. J. Appl. Crystallogr. 2005, 38, 1040. 373 (10) Baerlocher, C. EXTRACT. A Fortran program for the extraction 374 315 the closest. 1 of integrated intensities from a powder pattern; Institut fu¨r Kristallographie, 375 316 As deduced from the H static and MAS NMR spectra, the ETH: Zu¨rich, Switzerland, 1990. 376 317 crystal structure shows that the water molecules are, indeed, (11) Baerlocher, C.; Hepp, A. XRS-82. The X-ray RietVeld System; 377 318 isolated from one another. The closest approach is 4.66 Å (i.e., Institut fu¨r Kristallographie, ETH: Zu¨rich, Switzerland, 1982. 378 (12) Estermann, M. A.; Gramlich, V. J. Appl. Crystallogr. 1993, 26, 379 319 the length of the b axis) between water molecules coordinated + 1 396. 380 320 to adjacent K ions in the same channel. The H dipole coupling (13) Baerlocher, C.; McCusker, L. B.; Palatinus, L. Z. Kristallogr. 2007, 381 321 for such a separation is about 1.2 kHz, and its contribution can 222, 47. 382 322 be seen in the broadening of the powder pattern of the 1H NMR (14) Palatinus, L.; Chapuis, G. J. Appl. Crystallogr. 2007, 40, 786. 383 (15) Zhang, K. Y. J.; Main, P. Acta Crystallogr., A 1990, 46, 41. 384 323 spectrum shown in Figure 1b. In the present study, the sample ° (16) Palatinus, L.; van der Lee, A. J. Appl. Crystallogr. 2008, 41, 975. 385 324 was dehydrated up to 120 C under vacuum. No significant (17) International Tables for Crystallography; Wilson, A. J. C., Ed., 386 325 change was observed in either the 1Horthe11B MAS NMR, Kluwer Academic Publishers: Dordrecht, 1995, Vol. C; pp 684-692. 387 326 but a slight broadening was observed for 29Si MAS NMR (not (18) Grosse-Kunstleve, R. W.; McCusker, L. B.; Baerlocher, C. J. Appl. 388 Crystallogr. 1997, 30, 985. 389 327 shown). The framework stability in the absence of the crystalline 328 water would be an interesting issue for further investigation. JP903500Q 390