754

A multinuclear solid-state magnetic resonance and GIPAW DFT study of anhydrous calcium chloride and its hydrates

Can. J. Chem. Downloaded from www.nrcresearchpress.com by University of Ottawa on 10/03/11 For personal use only.

Cory M. Widdifield and David L. Bryce

Abstract: The group 2 metal halides and corresponding metal halide hydrates serve as useful model systems for understanding the relationship between the electric field gradient (EFG) and chemical shift (CS) tensors at the halogen nuclei and the local molecular and electronic structure. Here, we present a 35/37Cl and 43Ca solid-state nuclear magnetic resonance (SSNMR) study of CaCl2. The 35Cl nuclear quadrupole coupling constant, 8.82(8) MHz, and the isotropic chlorine CS, 105(8) ppm (with respect to dilute NaCl(aq)), are different from the values reported previously for this compound, as well as those reported for CaCl2·2H2O. Chlorine-35 SSNMR spectra are also presented for CaCl2·6H2O, and when taken in concert, the SSNMR observations for CaCl2, CaCl2·2H2O, and CaCl2·6H2O clearly demonstrate the sensitivity of the chlorine EFG and CS tensors to the local symmetry and to changes in the hydration state. For example, the value of diso decreases with increasing hydration. Gauge-including projector-augmented wave (GIPAW) density functional theory (DFT) calculations are used to substantiate the experimental SSNMR findings, to rule out the presence of other hydrates in our samples, to refine the hydrogen positions in CaCl2·2H2O, and to explore the isostructural relationship between CaCl2 and CaBr2. Finally, the 43Ca CS tensor span is measured to be 31(5) ppm for anhydrous CaCl2, which represents only the fifth CS tensor span measurement for calcium. Key words: solid-state NMR, chlorine, (CSA), electric field gradient.

35Cl, 37Cl, 43Ca,

CaCl2, polymorphism, hydrates, CQ, chemical shift anisotropy

Résumé : Les halogénures des métaux du groupe 2 et les hydrates d’halogénures métalliques sont des systèmes modèles qui permettent comprendre la relation entre le gradient du champ électrique (GCE) et les tenseurs du déplacement chimique (DC) au niveau du noyau d’halogène et la structure électronique et moléculaire locale. Dans ce travail, on présente les résultats d’une étude de résonance magnétique nucléaire à l’état solide (RMN-ES) du 35/37Cl et 43Ca du CaCl2. Les valeurs de la constante de couplage quadripolaire nucléaire du 35Cl, 8,82(8) MHz, et le DC isotrope du chlore, 105(8) ppm [par rapport au NaCl(aq) dilué], sont différentes de celles rapportées pour ce composé ainsi que les valeurs rapportés pour le CaCl2·2H2O. Les spectres RMN-ES du 35Cl sont aussi presentés pour le CaCl2·6H2O et, lorsqu’on les compare, les observations relatives aux spectres RMN-ES du CaCl2, du CaCl2·2H2O et du CaCl2·6H2O démontrent clairement la sensibilité du GCE du chlore et des tenseurs du DC sur la symétrie locale et aux changements dans l’état d’hydratation. Par exemple, les valeurs de diso diminuent avec une augmentation de l’hydratation. On a fait appel à des calculs selon la théorie de la fonctionnelle de la densité en utilisant une onde améliorée par un opérateur de projection comportant une jauge (« GIPAWDFT ») pour corroborer les observations faites par RMN-ES, pour éliminer la présence d'autres hydrates dans nos échantillons, pour affiner les positions des atomes d'hydrogène dans le CaCl2·2H2O et pour explorer la relation isostructurale entre le CaCl2 et CaBr2. Enfin, on a établi que la portée du tenseur de DC du 43Ca est de 31(5) ppm pour le CaCl2 anhydre, ce qui représente seulement la cinquième valeur mesurée pour l’anisotropie du tenseur du DC du calcium. Mots‐clés : RMN à l’état solide, chlore, 35Cl, mique (ADC), gradient du champ électrique.

37Cl, 43Ca,

CaCl2, polymorphie, hydrates, CQ, anisotropie du déplacement chi-

[Traduit par la Rédaction]

Introduction Solid-state nuclear magnetic resonance (SSNMR) spectroscopy is an important tool that is used to characterize struc-

ture and dynamics within a range of chemical systems. Recent developments in SSNMR1,2 have extended its applicability to a number of “exotic” nuclei. To obtain a more complete picture of a given solid-state system, however, it is

Received 2 November 2010. Accepted 22 December 2010. Published at www.nrcresearchpress.com/cjc on 3 May 2011. C.M. Widdifield and D.L. Bryce. Department of Chemistry and Centre for Catalysis Research and Innovation, University of Ottawa, 10 Marie Curie Private, Ottawa, ON K1N 6N5, Canada. Corresponding author: D.L. Bryce (e-mail: [email protected]). This article is part of a Special Issue dedicated to Professor Roderick E. Wasylishen in honour of his outstanding contributions to NMR spectroscopy, to chemistry in Canada, and to the training of NMR spectroscopists. Can. J. Chem. 89: 754 –763 (2011)

doi:10.1139/V11-009

Published by NRC Research Press

Can. J. Chem. Downloaded from www.nrcresearchpress.com by University of Ottawa on 10/03/11 For personal use only.

Widdifield and Bryce

often desirable to complement SSNMR data with information acquired using other methods. Recently, density functional theory (DFT) methods have been paired with ultrasoft pseudopotentials3 and the gauge-including projector-augmented wave (GIPAW) formalism4 to calculate magnetic shielding and electric field gradient (EFG) tensors at the nuclei in solid materials.4–6 These tensors are important SSNMR observables, and a number of recent studies have illustrated that magnetic shielding and EFG tensor calculations using GIPAW DFT can provide additional insights in the study of a variety of chemical systems.7–17 There is an increasing interest in “NMR crystallography”, which pertains to obtaining molecular-level and crystal structure information for solid materials using SSNMR data.18 Although there have been a number of exciting recent advances in this area of research,6,19–25 one cannot currently convert SSNMR data into a unique solid-state structure in a completely generalized manner. With a small amount of a priori knowledge, however, NMR crystallography methods have been applied to solve and refine the structures related to certain classes of materials, such as zeolites,26–28 which are sometimes subject to crystal twinning and typically do not possess enough long-range order to produce crystals that would be considered suitable for X-ray diffraction (XRD) studies. With the aid of computational modeling, SSNMR data acquired for spin-1/2 nuclei have been used to solve or refine structures of organic molecules such as b-L-aspartyl-Lalanine6 and thymol.22 We have recently contributed to this area by exploiting quadrupolar nuclei (i.e., nuclear spin quantum number, I > 1/2); for example, 79/81Br SSNMR data have been complemented with GIPAW DFT computations to refine the structure of the simple salt MgBr2.29 This work showed the pronounced sensitivity of the 79/81Br EFG tensor to minute structural changes. Quadrupolar nuclei possess a quadrupole moment (Q) which couples with the EFG at the nucleus to produce characteristic SSNMR spectral line shapes. These line shapes can usually be modeled using analytical simulation software, such as that developed by Wasylishen and co-workers, which treats this “quadrupolar interaction” (QI) as a second-order perturbation to the Zeeman states (the magnitude of which depends on the nuclear quadrupole coupling constant, CQ, and the quadrupole asymmetry parameter, hQ).30 Owing to the presence of the QI, SSNMR experiments using these probe nuclei should be able to offer additional structural restraints, relative to spin1/2 nuclei. A recent focus of our research group has been SSNMR spectroscopy of the quadrupolar halogen nuclides (35/37Cl, 79/81Br, and 127I).12,14,29,31–37 We have observed that the NMR spectra of these nuclei are very sensitive to the EFG produced by the ions and molecules of the surrounding environment. There exist two NMR-active chlorine nuclides (35Cl and 37Cl), both of which are quadrupolar (I(35/37Cl) = 3/2) and of moderately high abundance (75.779(46)% for 35Cl, 24.221(46)% for 37Cl).38 The Q associated with these nuclei are moderate (Q(35Cl) = –81.65(80) mb; Q(37Cl) = –64.35(64) mb);39 however, their magnetogyric ratios (g) are rather low (below that of 15N). Probe ringing and second-order QI line shape broadening are both reduced by performing the SSNMR experiments in as high an applied magnetic field (B0) as available. A variety of chlorine environments have been probed

755

using 35/37Cl SSNMR spectroscopy at “ultrahigh” B0, including (i) group 4 organometallic40,41 and group 13 inorganic materials,31 (ii) amino acid hydrochlorides,34,36 (iii) ionic liquids,42,43 (iv) group 2 metal chlorides,44 and (v) pharmaceutical hydrochlorides.45 As a result of these recent studies, a number of trends in the halogen SSNMR parameters have been observed and rationalized. For example, for the group 2 metal halides studied to date, it has been observed that sample hydration will (i) generally result in a decrease in the halogen isotropic chemical shift (CS) and (ii) nearly always lead to a decrease in the halogen CQ value.12,14,44 It was also noted, however, that the currently available 35Cl SSNMR data collected for CaCl2 indicate that this system does not follow the latter trend;46 rather, the data suggest that upon hydration the CQ value increases substantially. We therefore sought to carry out additional NMR studies on CaCl2, which included multiple B0 data acquisition and modern quantum chemical calculations, to confirm or refute the prior 35Cl SSNMR data. We present here the results of 35Cl SSNMR measurements for CaCl2 and CaCl2·6H2O. The 37Cl and 43Ca SSNMR spectra of anhydrous CaCl2 are also discussed. The results of GIPAW DFT computations on CaCl2 and its known crystalline hydrates (i.e., CaCl2·2H2O, (a/b/g)–CaCl2·4H2O, and CaCl2·6H2O) are discussed in an attempt to understand the prior 35Cl SSNMR measurements for CaCl2, to refine the hydrogen positions for CaCl2·2H2O, and to complement our experimental SSNMR data.

Experimental Sample preparation CaCl2 (99.99%) and CaCl2·6H2O (98%) were purchased from Sigma-Aldrich and received as powders. Sample purity for both compounds was confirmed by the manufacturer (see the Supplementary data, Additional Experimental). As CaCl2 is hygroscopic, it was stored and prepared for use under either dry N2 or Ar. CaCl2·6H2O was handled for short periods of time in a low humidity atmosphere to avoid further hydration. Prior to SSNMR experiments, the samples were powdered and then tightly packed into either 4 mm or 7 mm o.d. Bruker magic-angle spinning (MAS) ZrO2 rotors. Solid-state 35/37Cl and 43Ca NMR Data were acquired at the National Ultrahigh-field NMR Facility for Solids in Ottawa and at the University of Ottawa. Experiments at the Ultrahigh-field Facility used a standardbore Bruker AVANCE II spectrometer, which operates at B0 = 21.1 T (n0(1H) ≈ 900.08 MHz). For 35/37Cl SSNMR experiments at 21.1 T (n0(35Cl) = 88.189 MHz; n0(37Cl) = 73.408 MHz), home-built static probes (single-channel 7 mm or 4 mm HX) were used. Calcium-43 SSNMR experiments at the same field (n0(43Ca) = 60.575 MHz) employed a singlechannel 7 mm Bruker MAS probe. Experiments performed at the University of Ottawa used a wide-bore Bruker AVANCE spectrometer, which operates at B0 = 11.75 T (n0(1H) ≈ 500.13 MHz). All 35Cl SSNMR experiments at 11.75 T (n0(35Cl) = 49.002 MHz) used a 4 mm Bruker HX MAS probe. The 35/37Cl SSNMR spectra were referenced to 0.1 mol/dm3 NaCl in D2O at 0 ppm47 using solid KCl as a secondary reference (diso(KCl(s)) = 8.54 ppm).32 CalciumPublished by NRC Research Press

Can. J. Chem. Vol. 89, 2011

Can. J. Chem. Downloaded from www.nrcresearchpress.com by University of Ottawa on 10/03/11 For personal use only.

756

43 SSNMR spectra were referenced to a 2 mol/dm3 aqueous solution of CaCl2 (diso = 0 ppm).48 Chlorine and calcium pulse widths were established using the 35/37Cl and 43Ca SSNMR signals of solid KCl and saturated CaCl 2 (aq), respectively. Central transition (CT) selective (i.e., “solid p/2”) pulse widths were determined by scaling the p/2 pulse widths measured for the above standards by 1/ (I + 1/2) (i.e., 1/2 for 35/37Cl and 1/4 for 43Ca). All 35/37Cl SSNMR signals were acquired using the Solomon (i.e., “solid”) echo (i.e., p/2–t1–p/2–t2–acq)49–51 pulse sequence. Typical 35/37Cl SSNMR experimental parameters were as follows: CT-selective p/2 pulse width = 2.4– 3.2 µs; spectral window = 200–1000 kHz; t1 = 27.5– 50 µs; and 512–2048 complex time-domain data points. All 43Ca SSNMR signals were acquired using a singlepulse experiment (typical parameters: CT-selective p/2 ∼1.0 µs; spectral window = 20 kHz; 512–1024 complex time-domain data points). The pulse delays used, as well as the number of scans required for the collection of the spectra, may be found in the figure captions and the Supplementary data (Table S1). For 35Cl SSNMR experiments on CaCl2 at 11.75 T, variable offset cumulative spectrum (VOCS)52,53 data acquisition was used (offset = 110 kHz). After VOCS data acquisition, each component spectrum was processed as usual and combined in the frequency domain by co-addition. For CaCl2·6H2O, continuous-wave 1H decoupling was tested at both applied fields (n1(1H) ∼50 kHz at 11.75 T and ∼85 kHz at 21.1 T). For additional experimental details, see the Supplementary data, Table S1. Observed spectra were modeled using WSolids1, an analytical simulation software package developed by Eichele and Wasylishen.54 Quantum chemical calculations GIPAW DFT calculations employed CASTEP-NMR (v. 4.1) software,4,5,55,56 while the input files were generated using Materials Studio (v. 3.2.0.0) (Accelrys). Computations used either ultrasoft3 or “on-the-fly” pseudopotentials and the generalized gradient approximation. Further details pertaining to the relevant on-the-fly pseudopotentials may be found in the Supplementary data, Table S2. All geometry optimizations were carried out using the exchange-correlation (XC) functional of Perdew, Burke, and Ernzerhof (PBE),57,58 while the EFG and magnetic shielding tensor parameters were calculated using either the PBE XC functional or the XC functional of Perdew and Wang (PW91).59–63 Computed chlorine magnetic shielding tensor elements (sij) were transformed into CS tensor elements (dij) using the following procedure: the chlorine isotropic magnetic shielding value (siso) for NaCl was determined using a plane-wave cutoff energy (Ecut) of 1200 eV, a 4 × 4 × 4 Monkhorst–Pack k-point grid,64 and the same XC functional as the sample of interest. Using the above calculated siso value and the known chlorine chemical shift of solid NaCl relative to 0.1 mol/dm3 NaCl in D2O (i.e., diso(NaCl(s)) = –41.11 ppm),32 the siso value for 0.1 mol/dm3 NaCl in D2O could be calculated; hence, all computed sij values could be placed on an experimental d scale. The Ecut and k-point grid used for the GIPAW DFT calculations on each system are in the footnotes to Tables 1 and 2. For computed structure energies, structure references, optimized structure parameters, the pseudopotentials used,

and additional computational details, see the Supplementary data, Tables S2 and S3.

Results and discussion Experimental 35/37Cl and 43Ca SSNMR observations and analysis for CaCl2 The 35Cl EFG tensor parameters reported previously46 for CaCl2 appeared to be inconsistent with the trends observed for other group 2 metal halides,12,44 prompting us to reinvestigate this system using SSNMR spectroscopy. Chlorine-35 SSNMR spectra were obtained for CaCl2 at two applied fields (B0 = 11.75 and 21.1 T), and analytical modeling of all observed line shapes resulted in identical EFG and CS tensor parameters (Fig. 1). Additional 37Cl SSNMR experiments were performed at 21.1 T, and the spectra were modeled using the same parameters as those for the 35Cl SSNMR line shapes, after appropriate scaling of CQ by the Q(37Cl)/ Q(35Cl) ratio of 0.788 (Fig. 2). Extracted parameters are summarized in Table 1. It is immediately clear that the observed 35/37Cl SSNMR line shapes and EFG tensor parameters are inconsistent with the prior literature account for CaCl2.46 The presently measured CQ(35Cl) value of 8.82(8) MHz is more than twice as large as the values observed for other anhydrous and hydrated group 2 metal chlorides.44,46,69 This is primarily attributed to the arrangement of the Ca2+ ions within the first coordination sphere of each chloride, which forms a distorted trigonal plane.65 Our measured hQ(35/37Cl) value of 0.383(15) confirms the absence of high-symmetry rotational axes (i.e., Cn where n > 2), in support of the local symmetry for this site determined using XRD data, which is m.65 The observed chlorine chemical shift value (diso = 105 ± 8 ppm, or 146 ppm with respect to solid NaCl) is deshielded by 36 ppm relative to CaCl2·2H2O, which is consistent with the general trend observed for halogen chemical shift values upon hydration in group 2 metal halides.12 Experiments at 21.1 T were essential to characterize chlorine chemical shift anisotropy (CSA) and non-coincident EFG and CS tensor principal axis systems (PASs) in this sample. The prior 35Cl SSNMR study on CaCl2 was carried out under MAS conditions only and CSA information was not reported.46 Our measured chemical shift tensor span of 135 ± 15 ppm is by far the largest measured for a group 2 metal chloride system (72 ppm for CaCl2·2H2O being the second largest reported to date).44 The span value measured by modeling the 35Cl SSNMR spectrum is identical to the one extracted from the 37Cl SSNMR signal at 21.1 T (Fig. 2). As expected, the 35Cl SSNMR line shape of CaCl2 acquired at 11.75 T was not particularly sensitive to CSA effects, as nearly all the broadening could be attributed to the QI.70 The measured Euler angle values (these angles describe the relative orientation of the EFG and CS tensor PASs) for CaCl2 (i.e., a = 90° ± 20°, b = 90° ± 5°, g = 0° ± 5°) are consistent with the presence of mirror plane symmetry at the chloride anion, in agreement with prior X-ray data.65 Natural abundance 43Ca SSNMR experiments were also performed at B0 = 21.1 T to more completely characterize the sample. The resonance observed in the 43Ca MAS SSNMR spectrum (Fig. 3e) of CaCl2 (i.e., dpeak = 54.1 ± 0.2 ppm relative to 2 mol/dm3 CaCl2(aq)) cannot, unfortunately, be rigorously compared with that reported previously Published by NRC Research Press

Site label

CQ(35Cl) (MHz)a

EFG and chemical shift tensor parameters.

hQ

U (ppm)

k

diso (ppm)b

a (°)

b (°)

g (°)

Notes

—

Ref. 46 This study; CQ(37Cl) = 6.96(7) MHz X-ray structure from ref. 65

— — —

2.1(1) 8.82(8) 9.66

0.7(1) 0.383(15) 0.419

— 135(15) 200.9

— 0.0(3) 0.12

122(5) 105(8) 179.0

— 90(20) 90.0

— 90(5) 89.7

— — —

4.26(3) 6.69 4.43

0.75(3) 0.04 0.520

72(15) 141.4 140.7

0.6(2) 0.36 0.49

69(2) 114.0 109.3

90(10) — 118.3

82(5) — 80.8

—

0(5) 0.0 0(20) 2.7

Ref. 44 Ref. 44; non-optimized H positions This study; optimized H positions

Cl(1) Cl(2)

–3.70 5.97

0.405 0.944

73.6 77.9

0.39 –0.47

76.1 78.7

15.4 118.1

1.4 77.3

18.8 178.1

Optimized H positions Optimized H positions

Cl(1) Cl(2) Cl(3)

–3.75 2.16 3.97

0.947 0.828 0.776

123.5 83.6 125.6

0.70 0.59 0.45

92.9 78.9 104.6

183.1 135.5 65.3

3.2 73.6 82.6

269.3 199.5 357.7

Optimized H positions Optimized H positions Optimized H positions

—

3.59

0.149

77.1

0.12

48.5

111.1

47.7

190.6

Optimized H positions

— —

4.33(3) 5.48

57(3) 43.3

N/A N/A

90(7) 90.0

<0.01 0.000

40(8) 46.4

–1 –1

0(8) 0.3

— Neutron structure from ref. 66

Note: Measurement errors are within parentheses and parameters are defined as follows: CQ = eQV33/h; hQ = (V11 – V22)/V33, where |V11| ≤ |V22| ≤ |V33|; U ≈ d11 – d33; k ≈ 3(d22 – diso)/U; diso = (d11 + d22 + d33)/3, where d33 ≤ d22 ≤ d11; and a, b, and g describe the relative orientation between the EFG and CS tensor principal axis systems, which are disclosed here using the “ZYZ” convention67 and generated using EFGShield.68 All calculated results in this table used the PBE XC functional. Additional computational results and details can be found in the Supplementary data, Tables S2–S4. N/A, not applicable. a While CQ may take any real value, |CQ| is measured experimentally. Experimental signs are attributed based upon the GIPAW DFT results. b Experimental chlorine chemical shifts are relative to 0.1 mol/dm3 NaCl in D2O (diso(35/37Cl) = 0 ppm) and were established using solid KCl as a secondary reference (diso(35/37Cl) = 8.54 ppm). For the computational reference, NaCl(s), siso(35/37Cl) was calculated to be 975.06 ppm (4 × 4 × 4 k-point grid and Ecut = 1200 eV). Additional details can be found in the Experimental section. c Calculations for this compound used a 4 × 4 × 6 k-point grid and Ecut = 800 eV. d For the current study, calculations for this compound used a 4 × 3 × 2 k-point grid. Optimization of the hydrogen atomic positions used Ecut = 500 eV, and EFG and CS tensor parameter calculations used Ecut = 600 eV. e Calculations for this compound used a 4 × 4 × 3 k-point grid and Ecut = 650 eV. f Calculations for this compound used a 3 × 2 × 2 k-point grid and Ecut = 500 eV. g Calculations for this compound used a 4 × 3 × 3 k-point grid and Ecut = 650 eV.

757

Published by NRC Research Press

Can. J. Chem. Downloaded from www.nrcresearchpress.com by University of Ottawa on 10/03/11 For personal use only.

Experimental or computed Anhydrous CaCl2c Experimental Experimental Computed CaCl2·2H2Od Experimental Computed Computed a–CaCl2·4H2Oe Computed Computed b–CaCl2·4H2Of Computed Computed Computed g–CaCl2·4H2Og Computed CaCl2·6H2Oc Experimental Computed

35/37Cl

Widdifield and Bryce

Table 1. Experimental and GIPAW DFT-computed

758

Can. J. Chem. Vol. 89, 2011

Table 2. Experimental and GIPAW DFT-computed

Can. J. Chem. Downloaded from www.nrcresearchpress.com by University of Ottawa on 10/03/11 For personal use only.

Experimental or computed Experimental Computedd

CQ(43Ca) (MHz)a –0.95(20) –0.98

hQ 0.7(2) 0.541

43Ca

U (ppm) 31(5) 52.7

EFG and chemical shift tensor parameters for CaCl2. k –0.5(3) –0.24

diso (ppm)b 54.8(5) —e

a (°) 90.0c 90.0

b (°) 90.0c 89.4

g (°) 0.0c 0.0

Notes — X-ray structure from ref. 65

Note: Measurement errors are within parentheses. Parameter definitions can be found in the footnotes to Table 1. All calculated results in this table used the PBE XC functional. Additional computational results and details can be found in the Supplementary data, Tables S2, S3, and S5. a While CQ may take any real value, |CQ| is measured experimentally using SSNMR. Experimental signs are attributed based upon the GIPAW DFT results. b Experimental calcium chemical shifts are relative to dilute CaCl2(aq) (diso(43Ca) = 0 ppm) and were established using 2 mol/dm3 CaCl2(aq), which possesses an identical calcium CS as the primary standard. c Value assigned as a result of crystal symmetry and GIPAW DFT results. d Calculations for CaCl2 used a 4 × 4 × 6 k-point grid and Ecut = 800 eV. e Computed isotropic calcium magnetic shielding value: 1048.85 ppm.

Fig. 1. Analytical simulations (a, b, c, e), experimental static Solomon echo (d), and experimental static VOCS Solomon echo (f) 35Cl SSNMR spectra of powdered CaCl2, acquired at (d) B0 = 21.1 T (n0 = 88.189 MHz) and (f) B0 = 11.75 T (n0 = 49.002 MHz). In (a), the principal axis system orientations of the EFG and CS tensor frames are assumed to be equivalent (i.e., a = b = g = 0°), while in (b), CSA effects are ignored (i.e., U = 0 ppm). The best-fit spectra presented in (c) and (e) are simulated with the parameters given in Table 1. For (d), an optimized pulse delay of 10.0 s was used and 400 transients were collected, while for (f), a pulse delay of 1.0 s was used and 40 000 transients were collected for each of 6 transmitter frequencies (offset value used: 110 kHz).

Fig. 2. Analytical simulation (a) and experimental static Solomon echo 37Cl SSNMR spectrum (b) of powdered CaCl2, acquired at B0 = 21.1 T (n0 = 73.408 MHz). A pulse delay of 10.0 s was used and 960 transients were collected.

by Lin et al. (dpeak = 53.1 ± 2.0 ppm relative to saturated CaCl2(aq)).71 This is because the reference standard used in the earlier account (i.e., “saturated” CaCl2(aq)) shows a highly variable resonance position that depends on the preparation method (approx. ±10 ppm).48,72 Analytical modeling of the 43Ca MAS SSNMR line shape was used to establish approximate 43Ca EFG tensor parameters as well as the isotropic calcium CS for CaCl2: CQ(43Ca) = –0.95 ± 0.20 MHz, hQ = 0.7 ± 0.2, and diso = 54.8 ± 0.5 ppm, where the sign of CQ is based upon the results of GIPAW DFT computations (vide infra). Calcium-43 SSNMR experiments were also carried out under stationary conditions (Fig. 3b) and the parameters measured from the 43Ca MAS NMR spectrum were used to estimate the static line shape broadening that could be attributed to the 43Ca QI (Fig. 3c). It is clear that the QI does not dominate the line shape for this sample under static conditions at 21.1 T and that additional line broadening due to calcium CSA is present. Analytical line shape modeling of the static 43Ca SSNMR spectrum allows us to report a calcium CS tensor span of 31 ± 5 ppm (Fig. 3a), which represents the fifth measurement of calcium CSA in the literature73 and the first where oxygen atoms are not coordinated to the calcium. The calcium CS tensor span for CaCl2 is moderate, as calcium CS tensor spans range Published by NRC Research Press

Widdifield and Bryce

Can. J. Chem. Downloaded from www.nrcresearchpress.com by University of Ottawa on 10/03/11 For personal use only.

Fig. 3. Analytical simulations (a, c, d) and experimental single-pulse (b, e) 43Ca SSNMR spectra of powdered CaCl2, acquired at B0 = 21.1 T (n0 = 60.575 MHz) under (a, b, c) static and (d, e) MAS conditions. In (c), the effects due to calcium CSA are ignored (i.e., U = 0 ppm), while the simulation in (a) includes calcium CSA. The spectrum in (e) was acquired using a MAS frequency of 4 kHz. The experimental pulse delay for both (b) and (e) was 5.0 s. The spectrum in (b) is the result of the collection of 25 648 transients, while the spectrum in (e) is the result of the collection of 4313 transients.

from 8 ± 2 ppm in the calcite polymorph of CaCO3 to 70 ± 20 ppm in the vaterite polymorph of CaCO3.72 The remaining calcium SSNMR parameters for CaCl2 are summarized in Table 2. CaCl2·6H2O and other hydrates It is well known that CaCl2 is a strongly hygroscopic material, and as such it is widely used as a desiccant. In this section, we discuss the known experimental data for CaCl2 and selected hydrates to ensure that the spectral data above have been correctly interpreted. We have characterized CaCl2·6H2O using 35Cl SSNMR experiments at 11.75 and 21.1 T (Figs. 4 and 5), and analytical line shape modeling of both spectra yields identical parameters, which are summarized in Table 1. Our discussion of the 35/37Cl SSNMR data for the CaCl2 hydrates begins by noting that 35/37Cl SSNMR data were recently reported for CaCl2·2H2O44 and that this dihydrate possesses distinctly different EFG and CS tensor parameters compared with anhydrous CaCl2 (Table 1). To the best of our knowledge, two other hydrated forms of CaCl2 are stable under typical laboratory conditions: a–CaCl2·4H2O74 and CaCl2·6H2O.66 Additional polymorphic forms of the tetrahydrate (b–CaCl2·4H2O75 and g–CaCl2·4H2O76) may be prepared under normal conditions but appear to slowly convert to the a-form.77 As the hexahydrate is commercially available in a highly pure form, 35Cl SSNMR experiments were carried out on this material. By observing the 35Cl SSNMR line shapes associated with

759 Fig. 4. Analytical simulations (a, c) and experimental static Solomon echo (b, d) 35Cl SSNMR spectra of powdered CaCl2·6H2O, acquired at B0 = 11.75 T. The spectrum in (b) did not employ proton decoupling during signal acquisition, while the spectrum in (d) was acquired using proton decoupling. The simulation parameters used in (a) and (c) are identical, except for the addition of 1000 Hz of dipolar broadening in (a) (ad = bd = 0°, inferred from the neutron crystal structure), due to the expected 1H–35Cl distances of ∼2.27 Å. The experimental pulse delay for both (b) and (d) was 3.0 s. The spectrum in (b) is the result of the collection of 16 000 transients, while the spectrum in (d) is the result of the collection of 5616 transients.

CaCl2·6H2O (Figs. 4 and 5), it is clear that CaCl2·6H2O possesses a distinct line shape with respect to both anhydrous CaCl2 and CaCl2·2H2O. The measured 35Cl EFG tensor parameters (CQ(35Cl) = 4.33 ± 0.03 MHz, hQ < 0.01) strongly suggest that each chloride anion in CaCl2·6H2O is located on a C3 (or higher) rotational axis. The smaller value of CQ (35Cl) for this sample is consistent with earlier observations for group 2 metal halides, which nearly always exhibit smaller halogen CQ values upon hydration.12 It is also important to note that CaCl2·2H2O and all the polymorphs of CaCl2·4H2O place the chloride anions at positions lacking any local symmetry elements and hence the axial 35Cl EFG observed is unique to the hexahydrate. We also see a clear broadening of the discontinuities in Fig. 4b when 1H decoupling is not applied, which indicates that the sample is hydrated. The observed chlorine chemical shift for CaCl2·6H2O (diso = 57 ± 3 ppm) is smaller than that of both CaCl2 and CaCl2·2H2O, which is in agreement with earlier observations that the halogen shift value decreases (i.e., shielding increases) upon increasing hydration.12 The observed 35Cl SSNMR line shape of CaCl2·6H2O was best modeled when CSA effects were included (U = 40 ± 8 ppm, k = –1; also see Fig. 5). The observed chlorine U value is quite similar to that measured for other group 2 metal chloride hydrates.44 Based on the above discussion, it is clear that our anhydrous Published by NRC Research Press

760

Can. J. Chem. Downloaded from www.nrcresearchpress.com by University of Ottawa on 10/03/11 For personal use only.

Fig. 5. Analytical simulations (a, b) and experimental static Solomon echo (c) 35Cl{1H} SSNMR spectra of powdered CaCl2·6H2O, acquired at B0 = 21.1 T. The spectrum in (b) is the best fit to the experimental spectrum, while the spectrum in (a) does not include chlorine CSA (i.e., U = 0 ppm). A pulse delay of 5.0 s was used and 272 transients were collected.

CaCl2 sample cannot be CaCl2·2H2O or CaCl2·6H2O, and it is very unlikely to be a CaCl2·4H2O polymorph, owing to the observed large positive chlorine chemical shift value, the large CQ(35Cl) value, and the mirror site symmetry implied by the EFG and CS tensor interplay observed in the 35/37Cl SSNMR spectra of CaCl2. Quantum chemical computations GIPAW DFT quantum chemical calculations were carried out using the accepted crystal structures of CaCl2,65 CaCl2·2H2O,78 a–CaCl2·4H2O,74 b–CaCl2·4H2O,75 g– CaCl2·4H2O,76 and CaCl2·6H2O,66 although for the hydrates (except CaCl2·6H2O, whose structure was solved using neutron diffraction data) the H positions were optimized computationally (vide infra). The computed 35Cl EFG and chlorine CS tensor parameters are presented in Table 1, while more detailed computational results can be found in the Supplementary data, Table S4. Agreement between experimental and GIPAW DFTcomputed chlorine CS and EFG tensor parameters For nearly all of the relevant 35/37Cl NMR tensor parameters, excellent agreement between experimental observations and GIPAW DFT computations is seen for CaCl2 and CaCl2·6H2O. The calculated CQ(35Cl) values for these compounds are slightly overestimated relative to the experiment (∼1 MHz), which appears to be a general feature for chloride anions when using this computational method and pseudopotential.79 In addition, while there is a lack of quantitative agreement between the quantum chemical and experimental diso values, the pertinent experimental trend is reproduced computationally (i.e., as hydration increases, diso decreases). The quantitative disagreement in diso extends to the computational reference standard used in this study (i.e., NaCl(s)).

Can. J. Chem. Vol. 89, 2011

According to the absolute shielding scale established by Gee et al.,80 the experimental absolute shielding at the chlorine nuclei in solid NaCl is 1015 ± 4 ppm, while the present GIPAW DFT calculations provide a value of 975.06 ppm. Our data therefore suggest the presence of a systematic error in the calculation of siso values using the GIPAW DFT method and the current pseudopotential. We note that all calculated shift values reported in Table 1 used the computational shielding standard (i.e., siso(NaCl) = 975.06 ppm). There is also clear disagreement between the chlorine quantum chemical results for CaCl2 and the earlier 35Cl MAS NMR experimental observation:46 for example, the computed CQ(35Cl) is too large by roughly a factor of 4. Therefore, the quantum chemical result for CaCl2 provides independent evidence that the earlier 35Cl MAS NMR experimental measurements were either not correctly interpreted or were inadvertently carried out on a material other than CaCl2. The GIPAW DFT result for CaCl2·6H2O confirms the local symmetry (i.e., C3) for the chloride anions, as hQ = 0.0 and k = –1.0. It is also rather interesting to note the quantitative agreement between the calculated and experimental Euler angle values for CaCl2 and CaCl2·6H2O. While there typically exists very good agreement between experimental and computational values for the 35/37Cl EFG and magnetic shielding tensors in CaCl2 and CaCl2·6H2O, relatively poor agreement was observed for CaCl2·2H2O in a previous study.44 Inspection of the source data for the accepted crystal structure of CaCl2·2H2O reveals that the structure was generated using X-ray diffraction data, rather than neutron diffraction data, and hence the hydrogen positions are not precisely known. Out of all the group 2 metal chloride hydrates where 35Cl or 37Cl SSNMR data are available, CaCl2·2H2O represents the only structure for which neutron diffraction data are unavailable. The H positions in CaCl2·2H2O were therefore optimized using the GIPAW DFT method. Subsequent GIPAW DFT calculation of the EFG and magnetic shielding tensor parameters using this optimized structure yields markedly different tensor parameters as compared with the prior study (Table 1). Relative to the earlier computational results, there is a decrease in the computed CQ(35Cl) by about 34% (6.69 MHz to 4.43 MHz) and a substantial increase in hQ (0.04 to 0.520). Both of these significant changes result in better agreement between the experimental and computed EFG tensor parameter values. GIPAW DFT-computed chlorine EFG and CS tensor parameters for CaCl2·4H2O polymorphs Good agreement between the observed and computed 35/37Cl NMR tensor data was seen for the compounds above; hence, we explored whether GIPAW DFT computations would predict that chlorine SSNMR experiments could be used to differentiate between the a, b, and g polymorphs of CaCl2·4H2O. The differentiation of polymorphs using chlorine SSNMR has been demonstrated previously for chlorinecontaining pharmaceuticals,45 and as seen in Table 1, these experiments should also be able to distinguish the polymorphs of CaCl2·4H2O. Each polymorph should present a different number of magnetically unique chloride sites (ranging from 1 to 3). In addition, there exists significant disparity in the other relevant NMR parameters (notably hQ, which Published by NRC Research Press

761

Widdifield and Bryce

Can. J. Chem. Downloaded from www.nrcresearchpress.com by University of Ottawa on 10/03/11 For personal use only.

ranges from 0.149 in g–CaCl2·4H2O to above 0.776 for all three sites in b–CaCl2·4H2O). Are CaCl2 and CaBr2 isostructural? Based on prior XRD data, the ambient-condition structures of both CaCl2 and CaBr2 belong to the same space group and place the halogen anions at the same Wyckoff positions.65,81 The absolute CQ(81Br) value for CaBr2 is the largest amongst the group 2 metal bromides (CQ(81Br) = 62.8 ± 0.4 MHz).14 At the same time, the previously measured CQ(35Cl) value for CaCl2 was among the smallest measured relative to other group 2 metal chlorides (CQ(35Cl) = 2.1 MHz).46 The above sections have provided ample evidence that the earlier report was in error, likely because of the formation of a CaCl2 hydrate or the neglect to acquire spectral data using VOCS (or equivalent) methods. In addition, we now have reliable CQ(35/37Cl) measurements for CaCl2. It was recently shown by Wu and Terskikh82 that for a series of isostructural compounds, a linear relationship exists between the measured CQ values and the quantity Q(1 – g∞)/V, where g∞ is the Sternheimer antishielding factor83–86 and V is the unit cell volume. This finding has been extended by observations made for the 35/37Cl,44,46,69 79/81Br,14,29 and 127I12 nuclides within several isostructural series of group 2 metal halides and selected hydrates thereof. This linear relationship between CQ and Q(1 – g∞)/V is therefore useful for both confirming and predicting isostructural series within classes of analogous compounds. Although enough data do not exist in this case to construct a meaningful plot of CQ(X) vs. Q(1 – g∞)/V, we can compare the respective CQ(81Br)/CQ(35Cl) and [Q81Br(1 – g∞(Br))/V]/[Q35Cl(1 – g∞ (Cl))/V] ratios, which should be equivalent according to the above model if these two compounds are isostructural. Using the experimental SSNMR data for CaCl2 and CaBr2, the CQ(81Br)/CQ(35Cl) ratio is ∼7.12. Using additional known information, with the volume of the unit cells established using X-ray diffraction,65,81 the [Q81Br(1 – g∞(Br))/V]/[Q35Cl(1 – g∞(Cl))/V] ratio is calculated to be 5.19 (a 31.4% difference). This percent difference is comparable to what has been calculated previously for pairs of group 2 metal halide compounds that are understood to be isostructural; in fact, for the isostructural group 2 metal chlorides and bromides studied to date, it is observed that CQ(81Br)/CQ(35Cl) ratios range between 6.30 and 7.08, while the calculated [Q81Br(1 – g∞(Br))/V]/[Q35Cl(1 – g∞(Cl))/V] ratios for the same systems vary between 5.48 and 5.64.12 A likely contributor to the disagreement between the calculated [Q81Br(1 – g∞(Br))/V]/[Q35Cl(1 – g∞(Cl))/V] ratios and the experimental CQ(81Br)/CQ(35Cl) ratios is the uncertainty associated with the g∞ values: these values for the quadrupolar halogens are not known to high precision, unlike the group 1 metal atom g∞ values used in the study by Wu and Terskikh.82 At this time it is noted that, using the accepted structures for CaCl2 and CaBr2, the GIPAW DFT-calculated CQ(81Br)/CQ(35Cl) ratio is ∼7.16, which is in excellent agreement with our experimental observations (a difference of 0.56%). Likewise, upon consideration of the fine details associated with the CaCl2 and CaBr2 crystal structures, two points are clear: (i) the structure of CaCl2 is rather old (1935), and (ii)

the placement of the halogen anions is, in fact, not equivalent for the two structures (in CaBr2 the bromide is at x/a = 0.2636 ± 0.0029, y/b = 0.3417 ± 0.0026, z/c = 0, whereas in CaCl2 the chloride is at x/a = 0.275 ± 0.08, y/b = 0.325 ± 0.08, z/c = 0). It is therefore probable that this discrepancy in the halide positions within their respective crystalline lattices would contribute to the difference between the [Q81Br(1 – g∞(Br))/V]/[Q35Cl(1 – g∞(Cl))/V] ratio and the CQ (81Br)/CQ(35Cl) ratio. Indeed, it has been demonstrated recently that very small displacements of the halide anion (typically < 0.05 Å) can dramatically alter the calculated halogen CQ value for several similar systems.12,29 Therefore, while it is clear that the CQ/[Q(1 – g∞)/V] relationship appears to be of general use for group 2 metal halides, the precision of the halogen g∞ values is a limiting factor and the compounds must be rigorously isostructural to obtain exactly equal ratios.

Conclusions Solid-state 35/37Cl NMR experiments are sensitive probes of the local structure of inorganic materials and have allowed for differentiation between the hydrates of CaCl2. By performing chlorine SSNMR experiments in two applied magnetic fields, both the EFG and CS tensor parameters could be accurately measured in CaCl2 and CaCl2·6H2O. The measurement of these tensor parameters allowed us to correct the 35Cl SSNMR literature data for CaCl2 and illustrate that this species conforms to the established quadrupolar halogen NMR trends for group 2 metal halide systems (i.e., that both the halogen CQ value and the halogen chemical shift will decrease upon hydration). GIPAW DFT quantum chemical calculations have provided complementary data for the CaCl2 hydrates where experimental SSNMR data exist and predict that 35/37Cl SSNMR experiments should be able to distinguish the three known polymorphs of CaCl2·4H2O. It is seen that optimization of the H positions for CaCl2·2H2O, whose structure has been determined using Xray diffraction data, is necessary to obtain good agreement between the quantum chemical and experimental EFG tensor parameter values. By carrying out 43Ca SSNMR experiments at ultrahigh B0 (i.e., 21.1 T), we have been able to measure the fifth calcium CS tensor anisotropy, which is expected to prove useful in future SSNMR studies that correlate calcium CS tensor information to local structure and symmetry in other calcium-containing materials. Lastly, although it has been demonstrated that comparing CQ/[Q(1 – g∞)/V] ratios can provide evidence that a series of compounds are isostructural when using group 1 metal nuclei as probes, additional problems (i.e., uncertainty in halogen g∞ values and a requirement that the compounds be rigorously isostructural) present themselves when the quadrupolar halogen nuclei are used.

Supplementary data Supplementary data, including detailed 35/37Cl and 43Ca SSNMR experimental data acquisition parameters, GIPAW DFT quantum chemical calculations (pseudopotential files used, structure energies, structure references, crystal structure parameters used, and additional information), and GIPAW DFT-computed 35/37Cl and 43Ca EFG and magnetic shielding Published by NRC Research Press

762

Can. J. Chem. Vol. 89, 2011

tensor parameters, are available with the article through the journal Web site (www.nrcresearchpress.com/cjc). (18)

Can. J. Chem. Downloaded from www.nrcresearchpress.com by University of Ottawa on 10/03/11 For personal use only.

Acknowledgements D.L.B. would like to express his gratitude to Prof. Rod Wasylishen for his training and guidance throughout the years in the field of solid-state NMR spectroscopy and quantum chemistry. D.L.B. thanks the Natural Sciences and Engineering Research Council (NSERC) of Canada for funding. C.M.W. thanks NSERC for an Alexander Graham Bell CGS D2 scholarship. We are grateful to Dr. Victor Terskikh and Dr. Eric Ye for technical support. Access to the 900 MHz NMR spectrometer was provided by the National Ultrahighfield NMR Facility for Solids (Ottawa, Canada), a national research facility funded by the Canada Foundation for Innovation, the Ontario Innovation Trust, Recherche Québec, the National Research Council Canada, and Bruker Biospin and managed by the University of Ottawa (http://nmr900.ca). NSERC is acknowledged for a Major Resources Support grant.

(19)

(20)

(21)

(22)

(23) (24)

References (1) Lesage, A. Phys. Chem. Chem. Phys. 2009 11 (32), 6876. doi:10.1039/b907733m. (2) Ashbrook, S. E. Phys. Chem. Chem. Phys. 2009 11 (32), 6892. doi:10.1039/b907183k. (3) Vanderbilt, D. Phys. Rev. B 1990 41 (11), 7892. doi:10.1103/ PhysRevB.41.7892. (4) Pickard, C. J.; Mauri, F. Phys. Rev. B 2001 63 (24), 245101. doi:10.1103/PhysRevB.63.245101. (5) Profeta, M.; Mauri, F.; Pickard, C. J. J. Am. Chem. Soc. 2003 125 (2), 541. doi:10.1021/ja027124r. (6) Pickard, C. J.; Salager, E.; Pintacuda, G.; Elena, B.; Emsley, L. J. Am. Chem. Soc. 2007 129 (29), 8932. doi:10.1021/ ja071829h. (7) Johnston, K. E.; Tang, C. C.; Parker, J. E.; Knight, K. S.; Lightfoot, P.; Ashbrook, S. E. J. Am. Chem. Soc. 2010 132 (25), 8732. doi:10.1021/ja101860r. (8) Moudrakovski, I. L.; Alizadeh, R.; Beaudoin, J. J. Phys. Chem. Chem. Phys. 2010 12 (26), 6961. doi:10.1039/c000353k. (9) Hamaed, H.; Ye, E.; Udachin, K.; Schurko, R. W. J. Phys. Chem. B 2010 114 (18), 6014. doi:10.1021/jp102026m. (10) Middlemiss, D. S.; Blanc, F.; Pickard, C. J.; Grey, C. P. J. Magn. Reson. 2010 204 (1), 1. doi:10.1016/j.jmr.2010.01. 004. (11) Sutrisno, A.; Hanson, M. A.; Rupar, P. A.; Terskikh, V. V.; Baines, K. M.; Huang, Y. Chem. Commun. (Camb.) 2010 46 (16), 2817. doi:10.1039/b926071d. (12) Widdifield, C. M.; Bryce, D. L. J. Phys. Chem. A 2010 114 (40), 10810. doi:10.1021/jp108237x. (13) Hanna, J. V.; Pike, K. J.; Charpentier, T.; Kemp, T. F.; Smith, M. E.; Lucier, B. E. G.; Schurko, R. W.; Cahill, L. S. Chem.– Eur. J. 2010 16 (10), 3222. doi:10.1002/chem.200901581. (14) Widdifield, C. M.; Bryce, D. L. J. Phys. Chem. A 2010 114 (5), 2102. doi:10.1021/jp909106j. (15) Moudrakovski, I.; Lang, S.; Patchkovskii, S.; Ripmeester, J. J. Phys. Chem. A 2010 114 (1), 309. doi:10.1021/jp908206c. (16) Kibalchenko, M.; Yates, J. R.; Massobrio, C.; Pasquarello, A. Phys. Rev. B 2010 82 (2), 020202(R). doi:10.1103/PhysRevB. 82.020202. (17) Truflandier, L. A.; Boucher, F.; Payen, C.; Hajjar, R.; Millot,

(25)

(26) (27) (28) (29) (30)

(31) (32)

(33)

(34) (35) (36) (37) (38)

(39) (40)

Y.; Bonhomme, C.; Steunou, N. J. Am. Chem. Soc. 2010 132 (13), 4653. doi:10.1021/ja908973y. Harris, R. K. Crystallography & NMR: an Overview. In NMR Crystallography; Harris, R. K., Wasylishen, R. E., and Duer, M. J., Eds.; John Wiley & Sons: Chichester, UK, 2009; p 3. Webber, A. L.; Emsley, L.; Claramunt, R. M.; Brown, S. P. J. Phys. Chem. A 2010 114 (38), 10435. doi:10.1021/ jp104901j. Webber, A. L.; Elena, B.; Griffin, J. M.; Yates, J. R.; Pham, T. N.; Mauri, F.; Pickard, C. J.; Gil, A. M.; Stein, R.; Lesage, A.; Emsley, L.; Brown, S. P. Phys. Chem. Chem. Phys. 2010 12 (26), 6970. doi:10.1039/c001290d. Ganapathy, S.; Sengupta, S.; Wawrzyniak, P. K.; Huber, V.; Buda, F.; Baumeister, U.; Würthner, F.; de Groot, H. J. M. Proc. Natl. Acad. Sci. U.S.A. 2009 106 (28), 11472. doi:10. 1073/pnas.0811872106. Salager, E.; Stein, R. S.; Pickard, C. J.; Elena, B.; Emsley, L. Phys. Chem. Chem. Phys. 2009 11 (15), 2610. doi:10.1039/ b821018g. Seyfarth, L.; Senker, J. Phys. Chem. Chem. Phys. 2009 11 (18), 3522. doi:10.1039/b819319c. Ashbrook, S. E.; Cutajar, M.; Pickard, C. J.; Walton, R. I.; Wimperis, S. Phys. Chem. Chem. Phys. 2008 10 (37), 5754. doi:10.1039/b805681a. Harris, R. K.; Hodgkinson, P.; Pickard, C. J.; Yates, J. R.; Zorin, V. Magn. Reson. Chem. 2007 45 (S1), S174. doi:10. 1002/mrc.2132. Brouwer, D. H. J. Magn. Reson. 2008 194 (1), 136. doi:10. 1016/j.jmr.2008.06.020. Brouwer, D. H. J. Am. Chem. Soc. 2008 130 (20), 6306. doi:10.1021/ja800227f. Brouwer, D. H.; Kristiansen, P. E.; Fyfe, C. A.; Levitt, M. H. J. Am. Chem. Soc. 2005 127 (2), 542. doi:10.1021/ja043228l. Widdifield, C. M.; Bryce, D. L. Phys. Chem. Chem. Phys. 2009 11 (33), 7120. doi:10.1039/b911448n. Power, W. P.; Wasylishen, R. E.; Mooibroek, S.; Pettitt, B. A.; Danchura, W. J. Phys. Chem. 1990 94 (2), 591. doi:10.1021/ j100365a019. Chapman, R. P.; Bryce, D. L. Phys. Chem. Chem. Phys. 2009 11 (32), 6987. doi:10.1039/b906627f. Chapman, R. P.; Widdifield, C. M.; Bryce, D. L. Prog. Nucl. Magn. Reson. Spectrosc. 2009 55 (3), 215. doi:10.1016/j. pnmrs.2009.05.001. Widdifield, C. M.; Chapman, R. P.; Bryce, D. L. Chlorine, Bromine, and Iodine Solid-State NMR Spectroscopy. In Annual Reports on NMR Spectroscopy; Webb, G.A., Ed.; Elsevier Ltd.: Oxford, UK, 2009; Vol. 66, pp 195–326. doi:10. 1016/S0066-4103(08)00405-5. Chapman, R. P.; Bryce, D. L. Phys. Chem. Chem. Phys. 2007 9 (47), 6219. doi:10.1039/b712688c. Bryce, D. L.; Sward, G. D. J. Phys. Chem. B 2006 110 (51), 26461. doi:10.1021/jp065878c. Bryce, D. L.; Sward, G. D.; Adiga, S. J. Am. Chem. Soc. 2006 128 (6), 2121. doi:10.1021/ja057253i. Bryce, D. L.; Sward, G. D. Magn. Reson. Chem. 2006 44 (4), 409. doi:10.1002/mrc.1741. Coplen, T. B.; Böhlke, J. K.; De Bièvre, P.; Ding, T.; Holden, N. E.; Hopple, J. A.; Krouse, H. R.; Lamberty, A.; Peiser, H. S.; Révész, K.; Rieder, S. E.; Rosman, K. J. R.; Roth, E.; Taylor, P. D. P.; Vocke, R. D., Jr; Xiao, Y. K. Pure Appl. Chem. 2002 74 (10), 1987. doi:10.1351/pac200274101987. Pyykkö, P. Mol. Phys. 2008 106 (16–18), 1965. doi:10.1080/ 00268970802018367. Rossini, A. J.; Mills, R. W.; Briscoe, G. A.; Norton, E. L.; Published by NRC Research Press

Widdifield and Bryce

(41)

(42) (43)

Can. J. Chem. Downloaded from www.nrcresearchpress.com by University of Ottawa on 10/03/11 For personal use only.

(44) (45)

(46)

(47)

(48)

(49) (50) (51) (52)

(53) (54)

(55)

(56) (57) (58) (59)

(60)

(61)

Geier, S. J.; Hung, I.; Zheng, S.; Autschbach, J.; Schurko, R. W. J. Am. Chem. Soc. 2009 131 (9), 3317. doi:10.1021/ ja808390a. Hung, I.; Shetty, K.; Ellis, P. D.; Brey, W. W.; Gan, Z. Solid State Nucl. Magn. Reson. 2009 36 (4), 159. doi:10.1016/j. ssnmr.2009.10.001. Gordon, P. G.; Brouwer, D. H.; Ripmeester, J. A. J. Phys. Chem. A 2008 112 (49), 12527. doi:10.1021/jp808524h. Gordon, P. G.; Brouwer, D. H.; Ripmeester, J. A. ChemPhysChem 2010 11 (1), 260. doi:10.1002/cphc.200900624. Bryce, D. L.; Bultz, E. B. Chem.–Eur. J. 2007 13 (17), 4786. doi:10.1002/chem.200700056. Hamaed, H.; Pawlowski, J. M.; Cooper, B. F. T.; Fu, R.; Eichhorn, S. H.; Schurko, R. W. J. Am. Chem. Soc. 2008 130 (33), 11056. doi:10.1021/ja802486q. Sandland, T. O.; Du, L. S.; Stebbins, J. F.; Webster, J. D. Geochim. Cosmochim. Acta 2004 68 (24), 5059. doi:10.1016/j. gca.2004.07.017. Harris, R. K.; Becker, E. D.; Cabral De Menezes, S. M.; Granger, P.; Hoffman, R. E.; Zilm, K. W. Pure Appl. Chem. 2008 80 (1), 59. doi:10.1351/pac200880010059. Gervais, C.; Laurencin, D.; Wong, A.; Pourpoint, F.; Labram, J.; Woodward, B.; Howes, A. P.; Pike, K. J.; Dupree, R.; Mauri, F.; Bonhomme, C.; Smith, M. E. Chem. Phys. Lett. 2008 464 (1–3), 42. doi:10.1016/j.cplett.2008.09.004. Solomon, I. Phys. Rev. 1958 110 (1), 61. doi:10.1103/ PhysRev.110.61. Weisman, I. D.; Bennett, L. H. Phys. Rev. 1969 181 (3), 1341. doi:10.1103/PhysRev.181.1341. Kunwar, A. C.; Turner, G. L.; Oldfield, E. J. Magn. Reson. 1986 69, 124. doi:10.1016/0022-2364(86)90224-6. Massiot, D.; Farnan, I.; Gautier, N.; Trumeau, D.; Trokiner, A.; Coutures, J. P. Solid State Nucl. Magn. Reson. 1995 4 (4), 241. doi:10.1016/0926-2040(95)00002-8. Medek, A.; Frydman, V.; Frydman, L. J. Phys. Chem. A 1999 103 (25), 4830. doi:10.1021/jp990410d. Eichele, K.; Wasylishen, R. E. WSolids1: Solid-State NMR Spectrum Simulation Package, ver. 1.19.11; Universität Tübingen: Tübingen, 2009. Clark, S. J.; Segall, M. D.; Pickard, C. J.; Hasnip, P. J.; Probert, M. I. J.; Refson, K.; Payne, M. C. Z. Kristallogr. 2005 220 (5– 6), 567. doi:10.1524/zkri.220.5.567.65075. Yates, J. R.; Pickard, C. J.; Mauri, F. Phys. Rev. B 2007 76 (2), 024401. doi:10.1103/PhysRevB.76.024401. Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996 77 (18), 3865. doi:10.1103/PhysRevLett.77.3865. Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1997 78 (7), 1396. doi:10.1103/PhysRevLett.78.1396. Perdew, J. P. Unified Theory of Exchange and Correlation Beyond the Local Density Approximation. In Electronic Structure of Solids ’91: Proceedings of the 75th WEHeraeus-Seminar; Ziesche, P., Eschrig, H., Eds.; Akademie Verlag: Berlin, 1991; p 11. Perdew, J. P.; Chevary, J. A.; Vosko, S. H.; Jackson, K. A.; Pederson, M. R.; Singh, D. J.; Fiolhais, C. Phys. Rev. B 1992 46 (11), 6671. doi:10.1103/PhysRevB.46.6671. Perdew, J. P.; Chevary, J. A.; Vosko, S. H.; Jackson, K. A.;

763

(62) (63)

(64) (65) (66) (67) (68) (69) (70) (71) (72) (73) (74) (75) (76) (77) (78) (79) (80) (81)

(82) (83) (84) (85) (86)

Pederson, M. R.; Singh, D. J.; Fiolhais, C. Phys. Rev. B 1993 48 (7), 4978. doi:10.1103/PhysRevB.48.4978.2. Perdew, J. P.; Burke, K.; Wang, Y. Phys. Rev. B 1996 54 (23), 16533. doi:10.1103/PhysRevB.54.16533. Burke, K.; Perdew, J. P.; Wang, Y. Derivation of a Generalized Gradient Approximation: the PW91 Density Functional. In Electronic Density Functional Theory: Recent Progress and New Directions; Dobson, J. F., Vignale, G., and Das, M. P., Eds.; Plenum: New York, 1998. Monkhorst, H. J.; Pack, J. D. Phys. Rev. B 1976 13 (12), 5188. doi:10.1103/PhysRevB.13.5188. van Bever, A. K.; Nieuwenkamp, W. Z. Kristallogr. 1935 90, 374. Agron, P. A.; Busing, W. R. Acta Crystallogr. C 1986 C42 (2), 141. doi:10.1107/S0108270186097007. Arfken, G. B.; Weber, H. J. Mathematical Methods for Physicists, 6th ed. Academic Press: New York, 2005; p 202. Adiga, S.; Aebi, D.; Bryce, D. L. Can. J. Chem. 2007 85 (7–8), 496. doi:10.1139/v07-069. Stebbins, J. F.; Du, L. S. Am. Mineral. 2002 87 (2–3), 359. MacKenzie, K. J. D.; Smith, M. E. Multinuclear Solid-State NMR of Inorganic Materials; Pergamon: Amsterdam, 2002. Lin, Z.; Smith, M. E.; Sowrey, F. E.; Newport, R. J. Phys. Rev. B 2004 69 (22), 224107. doi:10.1103/PhysRevB.69.224107. Bryce, D. L.; Bultz, E. B.; Aebi, D. J. Am. Chem. Soc. 2008 130 (29), 9282. doi:10.1021/ja8017253. Bryce, D. L. Dalton Trans. 2010 39 (37), 8593. doi:10.1039/ c0dt00416b. Leclaire, A.; Borel, M. M. Acta Crystallogr. B 1979 B35 (3), 585. doi:10.1107/S0567740879004209. Leclaire, A.; Borel, M. M. Acta Crystallogr. B 1978 B34 (3), 900. doi:10.1107/S0567740878004288. Leclaire, A.; Borel, M. M.; Monier, J. C. Acta Crystallogr. B 1980 B36 (11), 2757. doi:10.1107/S0567740880009909. Shepelev, A. I.; Lyashenko, M. N.; Druzhinin, I. G. Dokl. Akad. Nauk SSSR 1950 75, 379. Leclaire, A.; Borel, M. M. Acta Crystallogr. B 1977 B33 (5), 1608. doi:10.1107/S0567740877006645. Chapman, R. P.; Hiscock, J. R.; Gale, P. A.; Bryce, D. L. Can. J. Chem. 2011 89. This issue. Gee, M.; Wasylishen, R. E.; Laaksonen, A. J. Phys. Chem. A 1999 103 (50), 10805. doi:10.1021/jp9925841. Brackett, E. B.; Brackett, T. E.; Sass, R. L. J. Inorg. Nucl. Chem. 1963 25 (10), 1295. doi:10.1016/0022-1902(63)803966. Wu, G.; Terskikh, V. J. Phys. Chem. A 2008 112 (41), 10359. doi:10.1021/jp8064739. Sternheimer, R. Phys. Rev. 1950 80 (1), 102. doi:10.1103/ PhysRev.80.102.2. Sternheimer, R. Phys. Rev. 1951 84 (2), 244. doi:10.1103/ PhysRev.84.244. Sternheimer, R. Phys. Rev. 1954 95 (3), 736. doi:10.1103/ PhysRev.95.736. Lucken, E. A. C. Sternheimer Shielding. In Nuclear Quadrupole Coupling Constants; Academic Press: London, 1969; p 79.

Published by NRC Research Press