Ni Solid-State NMR Spectroscopy - ACS Publications - American

Article pubs.acs.org/IC

Structural and Crystallographic Information from 61Ni Solid-State NMR Spectroscopy: Diamagnetic Nickel Compounds Peter Werhun and David L. Bryce* Department of Chemistry and Biomolecular Sciences and Centre for Catalysis Research and Innovation, University of Ottawa, 10 Marie Curie Private, Ottawa, Ontario K1N 6N5, Canada S Supporting Information *

ABSTRACT: Despite the significance of nickel compounds, NMR spectroscopy of the active nickel isotope 61Ni remains a largely unexplored field. While nickel(0) compounds have been studied by 61Ni NMR in solution, solid-state experiments have been limited to Knight shift studies of nickel metal and nickel intermetallics. In conjunction with an NMR study of their ligands and 61Ni relativistic computations, the first 61Ni solid-state NMR (SSNMR) spectra of diamagnetic compounds are reported here. Specifically, bis(1,5-cyclooctadiene)nickel(0) [Ni(cod)2], tetrakis(triphenylphosphite)nickel(0) [Ni[P(OPh)3]4], and tetrakis(triphenylphosphine)nickel(0) [Ni(PPh3)4] were studied. 61Ni SSNMR spectra of Ni(cod)2 were used to determine its isotropic chemical shift (δiso = 965 ± 10 ppm), span (Ω = 1700 ± 50 ppm), skew (κ = −0.15 ± 0.05), quadrupolar coupling constant (CQ = 2.0 ± 0.3 MHz), quadrupolar asymmetry parameter (η = 0.5 ± 0.2), and the relative orientation of the chemical shift and electric field gradient tensors. A solution study of Ni(cod)2 in C6D6 yielded a narrow 61Ni signal, and the temperature dependence of δiso(61Ni) was assessed (δiso being 936.5 ppm at 295 K). The solution is proposed as a secondary chemical shift reference for 61Ni NMR in lieu of the extremely toxic Ni(CO)4 primary reference. For Ni[P(OPh)3]4, 61Ni SSNMR was used to infer the presence of two distinct crystallographic sites and establish ranges for δiso in the solid state, as well as an upper bound for CQ (3.5 MHz for both sites). For Ni(PPh3)4, line shape fitting provided a δiso value of 515 ± 10 ppm, Ω of 50 ± 50 ppm, κ of 0.5 ± 0.5, CQ of 0.05 ± 0.01 MHz, and η of 0.0 ± 0.2. The study of Ni(PPh3)4, in particular, demonstrates the utility of 61Ni SSNMR given the lack of a previously reported crystal structure and transient nature of Ni(PPh3)4 in solution.

■

compounds14 and 55Mn SSNMR of solid manganese carbonyls15 (with the large quadrupole moment of 55Mn countering its otherwise high sensitivity) have both been previously reported. Here we provide, to our knowledge, the first 61Ni SSNMR characterizations of diamagnetic nickel complexes, both to demonstrate the viability of such studies and, coupled where required with indirect probing of the metal complexes using 31P and 13C NMR experiments and computational modeling, to give a more complete characterization of the compounds studied. The specific targets of this study were bis(1,5-cyclooctadiene)nickel(0) [Ni(cod) 2 ], tetrakis(triphenylphosphite)nickel(0) [Ni[P(OPh)3]4], and tetrakis(triphenylphosphine)nickel(0) [Ni(PPh3)4], which are depicted in Figure 1. As a note on the conventions used in this work, quadrupolar interaction tensor information is given by the quadrupolar coupling constant (CQ) and quadrupolar asymmetry parameter (η):

INTRODUCTION Nuclear magnetic resonance (NMR) spectroscopy is a universal component of the inorganic chemist’s toolkit, yet the vast majority of studies involve a handful of mostly spin 1/2 nuclides possessing some combination of high abundance, high sensitivity, and ubiquity in chemistry (a nonexhaustive list would include 1H, 13C, 19F, and 31P). Nonetheless, particularly with the availability of high-field NMR spectrometers, characterization of insensitive (including low-gyromagneticratio and/or strongly quadrupolar) nuclei opens up the possibility of directly studying centers of interest rather than relying solely on the surrounding nuclei to probe the chemical environment of a target site. To be sure, limited studies of insensitive nuclides have been utilized since the early days of NMR, but a great number of such nuclides, including 61Ni, remain understudied. Apart from Knight shift studies of alloys1,2 and metallic nickel,3,4 61Ni remains uncharacterized by NMR in the solid state and undercharacterized in solution.5−13 As will be discussed, this is largely due to the specific electronics and chemistry of nickel combined with its moderately unfavorable NMR properties, conspiring to leave 61 Ni behind while other insensitive nuclei have already provided insight into otherwise inaccessible realms of (in particular, solid-state) chemistry. For instance, solid-state 53Cr NMR (SSNMR) of diamagnetic chromium(0) and chromium(VI) © 2017 American Chemical Society

CQ =

eQV33 h

(1)

Received: June 16, 2017 Published: August 3, 2017 9996

DOI: 10.1021/acs.inorgchem.7b01536 Inorg. Chem. 2017, 56, 9996−10006

Article

Inorganic Chemistry

NMR characterization only of ligand nuclei. Some reviews on Pt NMR are referenced here17,18 for the interested reader and provide further examples of its useful applications. It is also worth noting that the 195Pt chemical shifts for several platinum(0) compounds, including Pt(cod)2,19 Pt(PMe3)4, Pt(PPh3)4,20 and Pt[P(OMe)3]4,21 in particular, have been previously reported but do not follow the same trends as 61Ni chemical shifts in solution for the analogous nickel compounds.7,10 The relatively high NMR sensitivity of 195Pt is in stark contrast to the other group 10 nuclides. As opposed to platinum, 105Pd has not been extensively explored by NMR because of its problematic quadrupole moment (about 4 times larger than that for 61Ni) and extremely low gyromagnetic ratio (about half that of the active nickel isotope). A limited body of work exists for this nuclide, with indirect coupling to phosphorus22 and Knight shift studies23,24 making up the majority. Only one solution NMR study appears to have been reported for 105Pd.25 61 Ni, despite having a smaller quadrupole moment and higher gyromagnetic ratio than 105Pd, is far from amenable to study by NMR, mostly because of its low natural abundance. Although not a group 10 element, in a discussion of the feasibility of 61Ni NMR experiments, it is worth examining 33S, which has nuclear properties (Table 1) broadly comparable to those of 61Ni. As is often the case, NMR spectroscopy for quadrupolar nuclei in highly symmetric sites can be relatively straightforward (such as for inorganic sulfides),26 whereas as in the case of S8,27 quadrupolar broadening may render NMR studies on low-symmetry species time-consuming and often highly impractical. Because of similarities between the nuclides, the results of 33 S NMR experiments can help shape expectations for nickel NMR (especially SSNMR) feasibility in high- and low-symmetry structures. 61 Ni NMR. Unlike 33S NMR, however, 61Ni NMR is complicated by the fact that most molecular nickel complexes and salts are paramagnetic and/or have square-planar geometries. Diamagnetic square-planar species occur, but the lower symmetry compared to that of the octahedral or tetrahedral coordination environments results in a degree of quadrupolar spectral broadening that renders impractical the acquisition of 61 Ni NMR spectra at present. However, several diamagnetic nickel species exist that are suitable candidates for NMR spectroscopy. Indeed, many have been previously characterized by NMR in solution, with one-bond and even two-bond J couplings routinely observed in the spectra of these compounds [with 1J(61Ni,31P) on the order of 150−500 Hz,6−8 and for Ni(PF3)4, 2J(61Ni,19F) = 17 Hz7]. These compounds are mostly tetrahedral or pseudotetrahedral, with isotropic chemical shifts (δiso) ranging from 937 ppm for Ni(cod)2 to −929 ppm for Ni(PF3)4.10 Less symmetric compounds have also been characterized by NMR in solution,10,11 including studies of 195

Figure 1. Three nickel(0) compounds studied in this work: (a) Ni[P(OPh)3]4; (b) Ni(PPh3)4; (c) Ni(cod)2.

η=

V11 − V22 V33

(2)

where |V11| ≤ |V22| ≤ |V33| are the principal components of the electric field gradient (EFG) tensor, Q is the nuclear quadrupole moment, and e is the fundamental charge. The Herzfeld−Berger chemical shift tensor notation16 is used, with the isotropic chemical shift (δiso) defined as δiso =

1 (δ11 + δ22 + δ33) 3

(3)

where δ33 ≤ δ22 ≤ δ11 are the principal components of the chemical shift tensor. The span (Ω) of a spectral line shape is given by σ − σ11 Ω = 33 ≈ δ11 − δ33 1 − σref (4) with σref being the reference magnetic shielding for the nuclide in question. The skew (κ) is

κ=

3(δ22 − δiso) Ω

(5)

NMR of Group 10 Elements. Before discussing various considerations specific to 61Ni NMR, it is useful to briefly survey the literature for related nuclides, specifically of group 10, in order to both assess the sensitivity of the 61Ni nuclide in context and gauge the theoretical difficulty of acquiring 61Ni NMR spectra. A comparison of the NMR-relevant properties of the group 10 elements is therefore provided in Table 1. Notably, 195Pt has the most favorable properties for NMR observation and is routinely employed for platinum-containing samples.17 Its broad chemical shift range (∼16000 ppm) allows for a relatively straightforward differentiation between isomeric structures and a high degree of sensitivity to ligand substitutions.10,17 The possibility of measuring the connectivity in nonequivalent platinum clusters10 is another reason why the direct detection of metal nuclei can be useful, as opposed to

Table 1. NMR Properties of 61Ni, Highlighting the Relatively Similar Characteristics of 61Ni to 33S and Dissimilarity Compared to the Other Group 10 NMR Nuclidesa I

natural abundance [%]

RC b

Q [×10−28 m2]c

γ [×107 rad s−1 T−1]d

Ξe

/2 /2 5 /2 1 /2

1.14 0.76 22.33 33.832

0.240 0.101 1.49 20.7

0.162 −0.0678 0.660 −

−2.3948 2.055685 −1.23 5.8385

0.08936051 0.07676000 0.04576100 0.21496784

nuclide 61

3

33

3

Ni S 105 Pd 195 Pt a

Values are adapted from ref 35. bReceptivity versus 13C at natural abundance. cNuclear quadrupole moment. dGyromagnetic ratio. eFrequency factor (relative to 1.000 for 1H). 9997


Article

Inorganic Chemistry di- and tricarbonylnickel complexes8,12,28 and, notably, the measurement of δiso for two trigonal-planar species.10,11 Knight shift studies of 61Ni in nickel intermetallics, while not directly comparable to the diamagnetic case, provide an example where nickel is found in both relatively high- and low-symmetry environments.1 In fact, significant physicochemical information could be collected for such materials using 61Ni NMR, either as a function of the aluminum stoichiometry, with a (lower-symmetry-environment-derived) baseline feature under a (higher-symmetry-site-originating) sharper peak for Ni1−xAlx with 0.492 ≤ x ≤ 0.520 or as a function of the temperature to observe phase transitions in NiTi.1 Full 61Ni SSNMR spectra were reported for Ni2Al3, Ni2Ga3, Ni2In3, NiAl3, Ni1−xGax with x = 0.52, 0.51, 0.50, 0.49, and 0.48, and Ni3Al4, NiIn, and NiTi with quadrupolar coupling constants (CQ) and quadrupolar asymmetry parameters (η) measured from spectral fitting for the latter three species (CQ = 5.0 MHz and η = 0.68 for Ni3Al4; CQ = 6.0 MHz and η = 0.67 for NiIn; CQ = 5.7 MHz and η = 0.30 for NiTi).1 As would be expected for intermetallics, the Knight shifts (Kiso) span a range of thousands of parts per million, from 7510 ppm for NiTi to 140 ppm for Ni3Al4. This is well outside the established diamagnetic chemical shift range of approximately −1000 to +1000 ppm for solution 61Ni NMR. Beyond their NMR suitability, nickel(0) compounds are used in a variety of syntheses as homogeneous catalysts, such that providing new means by which to study this set of compounds could help in assessing both the structure and its relationship to the function. Ni(cod)2, a homogeneous catalyst itself, is also the primary starting point for the synthesis of other nickel(0) catalysts.29,30 Ni[P(OPh)3]4 has been demonstrated to catalyze transfer hydrogenation,31 and Ni(PPh3)4 is used for, among other applications, catalytic cyanation32 and oxidation of phosphines to phosphine oxides.33 A full overview of nickel(0) chemistry is beyond the scope of this work, but a review published by Tasker et al.29 provides a good introduction to the wide breadth of homogeneous nickel catalysis. Understanding the chemistry and structure of nickel(0) complexes is therefore of pivotal importance for industrial applications and synthetic chemistry; to that end, this paper seeks to demonstrate the first steps toward the use of 61Ni NMR spectroscopy to explore these complexes in the solid state.

■

powder samples packed under an inert atmosphere in 7-mm-o.d. (outer diameter) borosilicate cells. 61Ni magic-angle-spinning (MAS) SSNMR spectra were acquired for Ni(PPh3)4 and Ni[P(OPh)3]4 by spinning at 8 kHz in high-speed (reduced-volume) 7-mm-o.d. zirconia rotors, while 31P MAS NMR spectra for these compounds were acquired in standard 7-mm-o.d. zirconia rotors at 4.7 T and standard 2.5-mm-o.d. zirconia rotors at 21.1 T. 61Ni MAS SSNMR spectra for Ni(cod)2 were acquired by spinning at 18 kHz in a 4-mm-o.d. reduced-volume zirconia rotor. For the Ni(cod)2 solid-state static 61Ni quadrupolar Carr−Purcell−Meiboom−Gill (QCPMG) experiments performed at 21.1 T, 1H decoupling was employed and found to slightly decrease the decay rate of the echo train, improving signal acquisition. Decoupling was not found to significantly affect the effective 61Ni relaxation rate in the other two compounds studied given the distance from the nickel center to the phenyl protons. Quadrupolar echo spectra (without 1H decoupling) were acquired for all compounds in the solid state at 21.1 T. For solution NMR studies, deuterated solvents were degassed using the freeze−pump−thaw method before drying as appropriate. The water content was assessed by relative peak integration in solution 1H NMR spectra and was below 300 ppm for acetone-d6, 50 ppm for acetonitrile-d3, and 30 ppm for benzene-d6. Ultimately, it was observed that compounds analyzed in solution were relatively insensitive to the water content over periods of days, although the compounds were extremely sensitive to oxygen. Solution samples consisted of saturated mixtures in 5 mm borosilicate NMR tubes, except for solution 61Ni NMR spectra, which were acquired from L-shaped 7-mm-o.d. borosilicate tubes custom-made for use in a solenoid coil, similar to the cell of Behringer and Blümel.8 All spectra were externally referenced: 31P SSNMR spectra were calibrated with the secondary reference triphenylphosphine (at −6 ppm for 4.7 T experiments) or ammonium dihydrogen phosphate (at +0.81 ppm for 21.1 T experiments); solution 31P NMR spectra were referenced relative to 85% H3PO4 (the primary reference at 0 ppm for 31 P); 13C and 1H referencing was performed relative to tetramethylsilane (TMS) in CDCl3. Because the IUPAC primary chemical shift reference for 61Ni, Ni(CO)4, is an extremely toxic, volatile liquid and given the lack of a readily available chemical shift reference in the literature, 61Ni NMR spectra were calibrated using the IUPAC unifiedscale convention; that is, the absolute frequency of Ni(CO)4 was calculated from the experimentally determined absolute frequency of 0.1% TMS in CDCl3 using the following relationship:35

Ξ x = (νx /νTMS) × 100%

(6)

where νx is the absolute frequency for the 0 ppm position in the NMR spectrum for nuclide x, νTMS is the absolute 1H frequency of TMS, and Ξx is the specific scaling factor for nuclide x (for 61Ni, see Table 1). Where 1H absolute referencing was impractical, chlorine (either 35/37Cl of NaCl solid or as a dilute solution) was used to determine the theoretical absolute frequency of the 1H TMS signal, as was previously done for 61Ni NMR.1 A saturated solution of Ni(cod)2 in C6D6 was found to produce an intense, narrow 61Ni signal that would lend itself well in the future to use as a secondary chemical shift reference instead of Ni(CO)4, given that a strong signal can be acquired in 20 min at 9.4 T and under 5 min at 21.1 T. The 61Ni chemical shift for this solution was referenced to the 1H signal of TMS and studied as a function of the temperature, with temperatures calibrated using neat ethylene glycol (according to the formula published by Merbach and coworkers36). 35/37 Cl NMR spectra were used to calibrate 61Ni pulse lengths because both chlorine nuclides are spin 3/2 with gyromagnetic ratios similar to that of 61Ni. Given the cubic symmetry of solid NaCl and quadrupolar averaging in solution, all π/2 pulses for 61Ni SSNMR were taken as half of the optimized chlorine pulse length, both to account for the behavior of spin 3/2 nuclei and to maximize the excitation bandwidth. 61Ni NMR peak shapes were optimized by varying the transmitter/receiver offset frequency and varying the relaxation delays (typical values used ranged from 0.2 to 2 s given the efficient relaxation of 61Ni, with 0.5 s giving a reproducible line shape at the shortest repetition time).

EXPERIMENTAL SECTION

NMR Spectroscopy. General experimental details are provided here, while specific parameters for particular spectra are provided in the relevant figures and captions. Ni(cod)2, Ni(PPh3)4 (with elemental composition, as determined by the manufacturer using an ethylenediaminetetraacetic acid titration, quoted as approximately 4−7% nickel by mass), and Ni[P(OPh)3]4 were purchased from SigmaAldrich and used without further purification. SSNMR studies were performed with these samples at 21.1, 11.7, 9.4, and 4.7 T, while solution NMR experiments were performed at 21.1, 14.1, and 7.0 T at the University of Ottawa and the National Ultrahigh-field NMR Facility for Solids. A table of the probes used is included in the Supporting Information. All samples were packed (for solids) or dissolved and transferred to borosilicate tubes (for solutions) under an inert atmosphere. Ni(PPh3)4 and Ni[P(OPh)3]4 were received as fine powders (red-brown and white, respectively). Two lots of Ni(cod)2 received had different physical characteristics: the first consisted of olive-gray crystals and was not ground before use or studied in solution because of the mechanochemical sensitivity of the compound; the second consisted of fine yellow crystals consistent with the literature34 and was ground before use and also characterized by solution NMR spectroscopy. Spectra were collected for stationary 9998


Article

Inorganic Chemistry

Table 2. Optimized Bond Lengths for Nickel Compounds As Computed in This Work Compared to Previous Computational and Experimental Literature Valuesa bond length [Å] compound

bond

ZORA-DFT

DFTb

Ni(CO)4 Ni(C2H4)2(PMe)

Ni−C Ni−C1 Ni−C2 Ni−P C1C2 Ni−C1 Ni−C2 C1C2 Ni−P Ni−P Ni−P Ni−P Ni−P Ni−P

1.817 1.997 2.000 2.159 1.406 2.114 2.125 1.395 2.103 2.160 2.148 2.298 2.148 2.162

1.829 2.008 2.011 2.177 1.412 2.121 2.140 1.401 2.107 2.180 2.154

Ni(cod)2

Ni(PF3)4 Ni(PMe3)4 Ni(PCl3)4 Ni(PPh3)4 Ni[P(OPh)3]4 Ni(PPh3)3

expt 1.825(2)c 2.000(2)d 2.015(2)d 2.238(1)d 1.383d 2.124(9)e 2.124(9)e 1.391(2)e 2.099(3)f

2.139(1)−2.156(1)g

a c

Atom indices are as in ref 11. The bond lengths are averaged for each bond in the molecule corresponding to the listed bond class. bFrom ref 11. From ref 50. dFrom ref 51. eFrom ref 41. fFrom ref 52. gFrom ref 53.

Computational Details. Relativistic spin−orbit zeroth-order regular approximation (ZORA) density functional theory (DFT) computations were performed for a series of nickel compounds by means of the Amsterdam Density Functional (ADF)37−39 computational chemistry program. These compounds were selected to expand upon a previous computational 61Ni NMR study [with Ni(cod)2, Ni(PMe3)4, Ni(PF3)4, Ni(PCl3)4, and Ni(C2H4)(PMe3) specifically being studied previously by computation]11 by including relativistic effects and results for a wider range of NMR parameters (particularly 61 Ni−31P J couplings and a broader characterization of the EFG tensor). The 61Ni NMR properties of Ni(PPh3)4, Ni[P(OPh)3]4, and Ni(PPh3)3 were studied for the first time in this work. Geometries were obtained for structures generated from an initial first-principles model using GaussView40 followed by a TZP set basis and BP86 functional DFT optimization computation using ADF. Crystal structure atomic coordinates were used for Ni(cod)241 but not for other compounds. Selected bond lengths were used to assess the geometry optimization for all compounds and are provided in Table 2. NMR parameters were computed using a TZP basis and B3LYP functional. The BP86 and B3LYP functionals have previously been demonstrated as reliable for nickel geometry optimization and NMR computations, respectively.11 Magnetic shieldings were converted to chemical shifts by either (i) direct calibration relative to Ni(CO)4 [i.e., using the calculated magnetic shielding of Ni(CO)4 as the reference shielding to determine the chemical shifts of the other compounds] or (ii) using the regression method demonstrated for 61Ni NMR by Bühl et al.11 Specifically, in the latter method chemical shifts were extrapolated from the fitting of the computed magnetic shielding values against the previously reported solution chemical shifts [with the exception of Ni(CO)4 itself].

Figure 2. Experimental and simulated 61Ni SSNMR spectra of static Ni(cod)2 at ν(1H) = 900 MHz. Fit parameters (Table 3) were determined by iterative fitting at multiple fields. For both experiments shown, equal numbers of transients were acquired at offsets of 0, +80, and +160 kHz relative to 0 ppm and the spectra summed to give the total line shape. A recycle delay of 0.5 s was used for both experiments. For the QCPMG experiment, 36k scans were acquired at each offset over a total of 17 h with a spikelet separation of 5 kHz. Proton decoupling slightly increased the length of the echo train. For the echo experiment, a modified quadrupolar echo was used without proton decoupling. Specifically, the second pulse was decreased from a π pulse to a π/2 pulse to maximize the effective excitation bandwidth, resulting in a π/2−τ−π/2 quadrupolar echo. For the echo, 152k scans were acquired at each offset with a 70 μs echo time for a total experiment time of 66 h. The inset shows experimental and simulated 61Ni NMR spectra of static Ni(cod)2 at ν(1H) = 400 MHz. Equal numbers of transients were acquired at offsets of 0, +36, and +66 kHz relative to 0 ppm and the spectra summed to give the total line shape. A total of 132k scans were acquired at each offset over a total of 60 h with a spikelet separation of 5 kHz.

■

RESULTS AND DISCUSSION Experimental Section. Ni(cod)2. Figures 2 and 3 depict what are believed to be the first characterizations of diamagnetic nickel compounds by 61Ni SSNMR spectroscopy. Figure 2 gives the spectrum for static, solid Ni(cod)2, with a quadrupolar echo experiment used at 21.1 T and a QCPMG sequence used at both 21.1 and 9.4 T. Figure 3 shows a 61Ni rotor-synchronized-echo MAS SSNMR experimental spectrum acquired at an 18 kHz spinning speed. The fit parameters for these spectra are summarized in Table 3. The experimental δiso of 965 ± 10 ppm is slightly greater than the solution value reported previously for this compound (δiso = 937 ppm)10 but reasonable given the difference in state (similar differences

between the solution and solid state can be seen for Ni[P(OPh)3]4). The span (Ω) of 1700 ± 50 ppm (compared 9999


Article

Inorganic Chemistry

rather than as an infinite collection of particles (which is presumed by theoretical treatments of powder SSNMR). Practically speaking, this means that, rather than obtaining a truly random distribution of crystallite orientations, there is a preference for a certain range of orientations in the sample, which will influence the effective anisotropic interactions observed for the bulk sample and cause distortions that cannot be accounted for by any single set of parameters for the overall spectrum. This effect has previously been observed in 35Cl SSNMR spectra of MgCl2, which forms platelets that can preferentially stack in the solid state under the pressure resulting from packing of the sample into a rotor.42 Nevertheless, between the use of multiple fields and inclusion of the MAS results, we are confident in the fit obtained for the yellow Ni(cod)2 material. However, it cannot be ruled out that magnetic field inhomogeneity may have also played some role in the nonstandard line shape, given that Ni(cod)2 decomposes slowly into nickel metal. This may explain why the tailing effect seen for the olive-gray sample can also be seen to a degree on the right-hand side of the echo experiment performed with the yellow sample. The echo experiment with the yellow lot was performed over a much longer time period than the QCPMG experiment (60 and 17 h, respectively) and was accompanied by some sample discoloration, acquiring a tan hue relative to the original yellow state, at the end of 60 h. Further, Ni(cod)2 has previously been characterized by single-crystal X-ray diffraction (unlike the other two compounds discussed in this work), which revealed that only one crystallographic site is present in the unit cell of this solid.41 The presence of only a single crystallographic site helps to justify the use of gas-phase calculations to assess the spectral fit, and as the data in Table 2 demonstrate, the experimental solidphase and computed gas-phase structures have similar characteristic bond lengths (Ni−C1 length of 2.114 Å and Ni−C2 length of 2.125 Å from ZORA-DFT in this work, 2.121 and 2.140 Å, respectively, for the DFT results from Bühl et al.,11 and 2.124(9) Å for both bonds from the experimental crystal structure41). Last, the identity of the yellow compound could be verified using solution NMR. The 1H NMR chemical shifts [1H NMR: δ 4.28 (br, CH), 2.05 (br, CH2)] are consistent with the literature,34 as is the 61Ni chemical shift (δiso = 936.5 ppm compared to the literature value of 937 ppm10). Solution 13C NMR also indicates the presence of the compound, with the cyclooctadiene ligands observed [13C NMR: δ 89.30 (CH), 30.49 (CH2)]. Beyond its usefulness in assessing the chemical identity of the yellow Ni(cod)2 material, the relative ease of 61Ni NMR in solution with this compound is promising for its use as a secondary chemical shift reference and as a means for pulse

Figure 3. Experimental 61Ni MAS NMR spectrum of Ni(cod)2 at ν(1H) = 900 MHz, with the isotropic peak and a quadrupolar parameter simulation in the inset. A rotor-synchronized π/2−τ−π/2 modified quadrupolar echo was used to acquire the spectrum at an 18 kHz spinning speed. Although this necessitated the use of a reducedvolume 4-mm-o.d. rotor, significantly reducing the amount of available sample in the coil, the 161 kHz static spectrum line width suggested that slower spinning speeds would not be able to effectively average the anisotropic interactions present. A total of 655k transients were acquired over 93 h with a recycle delay of 0.5 s.

to the ZORA DFT-computed value of 1306 ppm) suggests that for this compound the static line shape is dominated by chemical shift anisotropy, despite the not-insignificant quadrupolar coupling constant (CQ = 2.0 ± 0.3 MHz). With regard to the computationally determined parameters, the experimental CQ is smaller than the value determined both by this work (4.06 MHz) and by Bühl et al. (2.60 MHz).11 The experimental asymmetry parameter (η = 0.5 ± 0.2) is intermediate to the computational value of η = 0.36 obtained here and the computational results of Bühl et al. (η = 0.87). Experiments for Figures 2 and 3 made use of the yellow lot of Ni(cod)2, which gave slightly different results compared to the olive-gray material. As can be seen in the Supporting Information, the spectra obtained for the latter sample exhibited decreased intensity on the low-frequency edge of the central transition powder pattern relative to what would be predicted by simulation. This is attributed to a nonrandom crystallite orientational distribution in the sample stemming from the means by which it was prepared for characterization. Given that the sample was not homogenized by grinding prior to the experiment, as was done for the yellow material, the crystallites in the sample must be treated as a discrete number

Table 3. Experimentally Measured 61Ni NMR Parameters for the Chemical Species Studied in This Worka compound Ni(cod)2 Ni(PPh3)4 Ni[P(OPh)3]4 site 1 site 2

δiso [ppm] 965 (10) 515 (10)

Ω [ppm]

κ

|CQ| [MHz]

η

α [deg]

β [deg]

γ [deg]

1700 (50) 50 (50)

−0.15 (0.05) 0.5 (0.5)

2.0 (0.3) 0.05 (0.01)

0.5 (0.2) 0.0 (0.2)

65 (10) 0b

140 (5) 90 (45)

110 (5) 0c

−500 to −550 −590 to −635

0 to 3.5 0 to 3.5

a Errors are given in parentheses. Errors for Ni(cod)2 were obtained by iterative fitting at two fields; derivation of errors for Ni(PPh3)4 is presented in the Supporting Information. bSimulated line shape invariant to parameter variation. cSimulated line shape weakly sensitive to parameter variation but a conclusive range cannot be determined because of signal-to-noise limitations.

10000


Article

Inorganic Chemistry calibration, particularly at high fields. The narrow line width (ν1/2 of ∼50 Hz at 9.4 T and ∼100 Hz at 21.1 T) enables the acquisition of good signal-to-noise spectra in reasonable time periods with a high degree of precision in the isotropic chemical shift. A 61Ni NMR nutation curve using this sample produced results similar to those obtained from the 35/37Cl resonance of NaCl (solid or solution), increasing confidence in the use of chlorine to calibrate the 61Ni NMR pulse lengths. The chemical shift was observed to be strongly and linearly temperaturedependent; a plot is provided in the Supporting Information, with the final equation of the fit being

obtained by 61Ni MAS NMR, results in too many unknowns to allow for a reliable fitting. The solution 61Ni chemical shift for Ni[P(OPh)3]4 was reported previously,10 but here we are able to also provide an estimate of the 61Ni−31P J coupling as seen for other Ni−Pbond-containing species in the literature (Figure 5). There is a

δiso[Ni(cod)2 , 61Ni] = (0.4274 ± 0.0063)T + (810.3 ± 1.9)

(7)

61

On the basis of this equation, the Ni chemical shift at 295 K is 936.5 ± 2.7 ppm, with the error relative to the IUPAC and Ni(CO)4 referencing methods and taking into account the temperature dependence of the chemical shift. Ni[P(OPh)3]4. Experimental solid-state 61Ni NMR spectra of Ni[P(OPh)3]4 are provided in Figure 4 (both static and MAS at

Figure 5. Experimental 61Ni NMR spectrum of Ni[P(OPh)3]4 in benzene-d6 at ν(1H) = 900 MHz. Note the possible presence of the expected Ni−P pentet, with 1J(61Ni,31P) = 400 ± 20 Hz suggested, similar to that observed for related species (specifically 398 Hz for Ni[P(OMe)3]4).4 A total of 66k scans were acquired over 10 h with a recycle delay of 0.5 s.

difference of about 10 ppm in the reported δiso values (with −587 ppm measured here and −576 ppm reported in the literature10), but given the solvent sensitivity of the 61Ni isotropic chemical shifts11 and the lack of a solvent reported with the literature chemical shift,10 the discrepancy is not especially concerning. In Figure 5, the central three peaks of the expected pentet fine structure appear to be visible (although S/ N considerations prevent conclusive interpretation), with the suggested one-bond J coupling value of 1J(61Ni,31P) = 400 ± 20 Hz within the same range as that found for the chemically similar Ni[P(OMe)3]4 (both experimentally at 398 Hz7 and computationally at 396 Hz) and equivalent within error to the value predicted from DFT calculations on Ni[P(OPh)3]4 itself (408 Hz). Solution NMR experiments provide further evidence that the two signals observed in the solid state are associated with two crystallographically distinct sites rather than multiple chemical species. Doping experiments with triphenylphosphite were able to clearly separate and identify free and bound ligand peaks, with the 31P NMR spectrum in particular demonstrating the presence of only one Ni[P(OPh)3]4 chemical environment in the sample upon dissolution of the solid in C6D6 (Figure 6). This, combined with the signal observed in the solution 61Ni NMR spectrum, strongly suggests that the two peaks observed in the solid state stem from multiple crystallographically inequivalent molecules in the asymmetric unit rather than from multiple chemical species. In theory, the 61Ni−31P J coupling could perhaps be seen in the solution 31P NMR spectrum, but as a result of the quadrupolar nature and extremely low natural abundance of 61Ni, 61Ni−31P satellite couplings are almost always broadened into the baseline.8 It is worth mentioning that the 31P−13C J coupling observed for the unbound triphenylphosphite cannot be seen in the bound ligand; even accounting for the increased line widths observed for the latter, the J coupling must be significantly smaller in order to not be detected (Figure 6). This likely stems from the effects of electron delocalization from the P−O−C

Figure 4. Experimental 61Ni NMR spectra of Ni[P(OPh)3]4 at ν(1H) = 900 MHz under static and MAS conditions. Under sample spinning, the broad static spectrum narrows to two sites with integration of 2.1 to 1.0. Given that only one site appears in solution (Figure 5), this would appear to stem from multiple crystallographic sites rather than multiple species. For static at high field, 174k scans were acquired over 20 h using a π/2−τ−π quadrupolar echo with an echo time of 70 μs and a recycle delay of 0.5 s. For MAS NMR, 436k scans were acquired over 3 days with a recycle delay of 0.5 s at an 8 kHz spinning speed.

8 kHz). In the case of Ni[P(OPh)3]4, only limited line shape fitting could be performed because of the overlap between multiple sites, with two sites clearly resolved under MAS yielding relative integrations of 1.0 and 2.1. Keeping in mind that the integrations are approximate given the relatively low signal-to-noise ratio, in the absence of impurities, this would suggest the presence of multiple crystallographic sites. This is further supported by an inability to fit the MAS NMR line shapes adequately with anything but a two-site model. Upper bounds of CQ and ranges for δiso, as determined by MAS NMR, are included in Table 3. For both sites, CQ was estimated to be less than 3.5 MHz using a two-site fitting and variable line broadening, while the δiso values could only be given as a range (δiso = −500 to −550 for one site and −590 to −635 for the other) as the fitted isotropic chemical shift varies with the simulated quadrupolar coupling constant because of the second-order quadrupolar shift. The static spectrum could not be fit because of a lack of crystallographic information, which, coupled with the uncertainty in the CQ and δiso values 10001


Article

Inorganic Chemistry

Figure 7. Experimental and simulated 61Ni NMR spectra of static Ni(PPh3)4 at ν(1H) = 900 MHz. Fit parameters (Table 3) were determined by analysis at one field, using the shape of the satellite transitions as a guide; a demonstration of this process is available in the Supporting Information. A total of 174k scans were acquired over 20 h using a π/2−τ−π quadrupolar echo with an echo time of 70 μs and a recycle delay of 0.5 s.

8), while not exemplifying the satellite transition spinning sideband manifold that would be expected at 8 kHz (likely due to a Figure 6. Solution 31P (top) and 13C (bottom) NMR spectra of Ni[P(OPh)3]4 in benzene-d6 with the doped experiments including dissolved triphenylphosphite. Given that only one peak was observed in the solution 31P NMR spectrum for this compound, it is unlikely that multiple chemical species exist in the starting solid. The asterisk denotes the solvent carbon 1:1:1 triplet.

region to the Ni−P−O moiety. In fact, the effects of ligand association in the case of triphenylphosphite have been previously reported. Single substitution from Ni(CO)4 to Ni(CO)3[P(OPh)3] results in J coupling values of 2.9 and 5.0 Hz for the one- and two-bond couplings compared to 3.4 and 6.9 Hz in the free ligand,43 the decrease being consistent with a J coupling too small to be resolved in the case of fully substituted Ni[P(OPh)3]4 (though, from the spectra, unambiguously below 2 Hz). Ni(PPh3)4. 61Ni static and MAS SSNMR spectra collected for Ni(PPh3)4, shown in Figure 7, reveal a rather narrow (30 kHz central transition line width) line shape flanked by two “horns”, corresponding to the central (m = 1/2 ↔ −1/2) and satellite (m = 3/2 ↔ 1/2 and m = −1/2 ↔ −3/2) transitions, respectively. As shown in the Supporting Information, the static line shape was fit (Table 3) at one field using the satellite transitions, with the quadrupolar coupling constant (CQ = 0.05 ± 0.01 MHz) of similar magnitude to that computationally predicted for Ni(PMe3)4 (0.15 MHz). The ZORA-DFT computed span of 132 ppm for Ni(PPh3)4 is similar to the experimental span of 50 ± 50 ppm. Given that the 61Ni chemical shift in solution has not previously been reported for this compound, computational and experimental results for the model Ni(PMe3)4 and computed values for Ni(PPh3)4 itself, in addition to the trend found in the chemical shifts upon substitution from Ni(CO)4 to Ni(CO)3(Ph) to Ni(CO)2(Ph)2,8 were used to guide the search for the general peak location in the solid state and, in fact, match what would be expected [with isotropic chemical shifts of 24.2, 90.7, and 515 ppm for Ni(CO)3(Ph), Ni(CO)2(Ph)2, and Ni(PPh3)4, respectively]. The single-pulse 61 Ni MAS NMR spectrum acquired for this compound (Figure

Figure 8. Experimental MAS and static (see Figure 7) 61Ni NMR spectra of Ni(PPh3)4 at ν(1H) = 900 MHz. The single (relatively) narrow peak does not suggest the presence of multiple crystallographically distinct Ni(PPh3)4 sites in the solid phase. For the MAS NMR spectrum, a total of 461k scans were acquired over 67 h with a recycle delay of 0.5 s at an 8 kHz spinning speed.

low signal-to-noise ratio), is still useful in that it does not suggest that the signal in the static spectrum originates from multiple crystallographic sites. Interestingly, despite giving narrower static solid-state line shapes than either of the other two compounds, Ni(PPh3)4 gives much weaker NMR signals for each nuclide studied over the same acquisition time. This is likely a result of the chemistry of this compound; Ni(PPh3)4 has been demonstrated to form an equilibrium with the trigonal-planar species and free ligand in solution (Scheme 1).44−46 Given that the purity of the compound as provided was only quoted in terms of its nickel content, it is entirely possible and indeed, based on the 31P MAS NMR spectra acquired (provided in the Supporting Information), likely that the solid used was a mixture of the trigonal-planar Ni(PPh3)3 and tetrahedral Ni(PPh3)4. As can be seen from solution 31P NMR, Scheme 1

10002


Article

Inorganic Chemistry

complex in the solid state, as opposed to the solid being solely the trigonal-planar species with the extra triphenylphosphine ligand incorporated as part of the crystal lattice. Between the 31 P MAS and 61Ni static NMR experiments and the color of the solid material, it would appear that both of the hypotheses put forward by Tolman et al. regarding the nature of this solid were correct:44 solid Ni(PPh3)4 appears to be a mixture of bound ligand and the tetra- and tricoordinated species, with their suggestion of crystal packing forces stabilizing Ni(PPh3)4 being a likely explanation for the presence of the otherwise unstable complex. This would also explain the difficulty in characterization by X-ray crystallography, specifically explaining the apparent disorder in the crystal. It is worth noting that while the quadrupolar coupling constant is the predominant indicator of the presence of Ni(PPh3)4, its value does not necessarily agree with the DFT computations (Table 4). While the experimental CQ for

the chemical shift of the tetrahedral signal is observed at lower frequency relative to the trigonal-planar -(PPh3)3 signal (Figure 9). Correspondingly, the number of signals present in the 31P

Table 4. Comparison of the Single-Molecule ZORA-DFT Computational EFG Results with Previously Reported GasPhase Computational Work by Bühl et al.11 and Experimental Solid-State Values from This Worka compound Ni(CO)4 Ni(cod)2 Ni(PMe3)4 Ni[P(OMe)3]4 Ni(PF3)4 Ni(PCl3)4 Ni(C2H4)2 (PMe3) Ni(PPh3)4 Ni[P(OPh)3]4 Ni(PPh3)3

Figure 9. Solution 31P (top) and 13C (bottom) NMR spectra of Ni(PPh3)4 in benzene-d6, acetone-d6, and acetonitrile-d3, with doped samples including dissolved triphenylphosphine. The 31P inset highlights the broad triphenylphosphine peak present in acetonitrile; the relative integration of 1.0−3.0 for the narrow peaks present in this solvent is evidence that the broad peak in C6D6 [corresponding to Ni(PPh3)4 in fast exchange] is completely converted to Ni(PPh3)3 and PPh3 in acetonitrile. In the 13C spectrum, the asterisk denotes the solvent 1:1:1 triplet of benzene. The presence of broadened peaks in both solvents indicates the presence of a fast-exchanging species [Ni(PPh3)4], but the sharper peaks present in acetone indicate the presence of a significant amount of Ni(PPh3)3 and triphenylphosphine.

|CQ| [MHz]b 2.0 (0.3)

0.05 (0.05) 0−3.5

calcd |CQ| [MHz]b 0.09 2.60,c 4.06 0.16 1.65 0.04 0.02 38.0c 30.9 5.58 4.37 21.0

η 0.5 (0.2)

0.0 (0.2)

calcd η 0.19 0.87,c 0.36 0.90 0.24 0.60 0.40 0.20c 0.27 0.08 0.94 0.09

a All values are determined for the 61Ni site in each compound. Computational details can be found in the main text. Errors are given in parentheses for the experimental values. bCQ can take the value of any real number, but only the magnitude is measurable by NMR spectroscopy. cFrom ref 11.

Ni(PPh3)4 (0.05 ± 0.01 MHz) is much smaller than that predicted by computation for a model of Ni(PPh3)4, CQ is very close to that predicted for Ni(PMe3)4 (0.15 MHz) and is much too small to be attributable to Ni(PPh3)3, which should be broadened beyond the observed spectral window (with a CQ value of around 20 MHz predicted, corresponding to a 640 kHz central transition line width at 21.1 T for 61Ni NMR). The reason why the quadrupolar coupling is overestimated for Ni(PPh3)4 in the computations is likely due to the fact that the models use gas-phase structures. Given that Ni(PPh3)4 is unstable outside of the solid state, the gas-phase structure is unlikely to be representative. Indeed, the bond lengths predicted by DFT support this suggestion (Table 2). Specifically, the Ni−P bond length in Ni(PPh3)4 appears to be significantly longer than for any of the other related nickel compounds modeled, particularly Ni(PMe3)4 and Ni(PPh3)3. In fact, while Ni(PMe3)4, Ni(PPh3)3, and Ni[P(OPh)3]4 have essentially identical Ni−P bond lengths (2.160, 2.162, and 2.148 Å, respectively), Ni(PPh3)4 has a Ni−P bond length of 2.298 Å, about 0.15 Å longer than that of Ni[P(OPh)3]4 (keeping in mind that, although the computations were run

MAS NMR spectra and their chemical shifts would strongly suggest the presence of both of these particular chemical species in the solid state as well, accounting for the low 61Ni NMR signal intensity per unit volume compared to the other compounds: there is simply less Ni(PPh3)4 as a total proportion of the solid (with the tricoordinated species being unobservable by 61Ni NMR). Further evidence of the presence of Ni(PPh3)3 is provided by previous UV−vis characterizations, which have demonstrated that the Ni(PPh3)3 form is red-orange,44 explaining the overall red-brown and red colors of the solids and solutions, respectively. In fact, it has been suggested by Tolman et al. that the solid form exists only as Ni(PPh3)3·PPh3, with no presence of the tetracoordinated species at all.44 While single-crystal X-ray structures have not previously been obtained for this solid because of crystal disorder,45 NMR spectroscopy depends only on local order. The 31P MAS NMR spectra may suggest a complicated mixture of products, but alone they are far from conclusive, and it is the 61Ni NMR spectra presented here that provide strong evidence for the existence of the high-symmetry Ni(PPh3)4 as a discrete 10003


Article

Inorganic Chemistry

Table 5. Comparison of the Single-Molecule ZORA-DFT-Computed Isotropic Chemical Shift and J Coupling Results with Previously Reported Solution Experimental Values and Gas-Phase Computational Worka compound

δiso [ppm]

Ni(CO)4 Ni(cod)2 Ni(PMe3)4 Ni[P(OMe)3]4 Ni(PF3)4 Ni(PCl3)4 Ni(C2H4)2(PMe3) Ni(PPh3)4 Ni[P(OPh)3]4 Ni(PPh3)3

937b,c 40c −742g −929g 267g −866d 515b,I −587,b −576c

calcd δiso [ppm] 0 [defined] 802d 23d −738d 245d −937d

347e −240e −684e −879e −18e −988e 103e −618e −562e

561f −146f −680f −916f 121f −1048f 267f −602f −534f

J(61Ni,31P) [Hz]

calcd J(61Ni,31P)b [Hz]

285g 398h 482h 450h

246 396 413 385 227 243 408 326

400b

a

All values are determined for the 61Ni site in each compound. Computational details can be found in the main text. Boldface compounds were studied by the experiment in this work. bFrom this work. cFrom ref 10. dFrom ref 11. eFrom the regression method. fFrom direct calibration versus Ni(CO)4 magnetic shielding. gFrom ref 13. hFrom ref 7. IFor this compound, the isotropic chemical shift is for the solid.

obtained from the computed magnetic shielding values either by using a correlation method or by using Ni(CO)4 as the reference. It is worth noting that, in general, the chemical shifts calculated by calibrating versus Ni(CO)4 are significantly closer to the experimental values from solution (including literature values7,11,12) and the solid state than those obtained using the correlation method (a plot of the correlation between the experimental and calculated chemical shifts for the two methods is provided in the Supporting Information). As per Bühl et al., Ni(CO)4 is excluded when adopting the correlation method because they found that with the B3LYP functional paramagnetic contributions to the nickel magnetic shielding tensor in Ni(CO)4 are significantly overestimated relative to other nickel compounds (which was also observed in this work).11 Overall, it appears the computational−experimental correlation is good for low-frequency chemical shifts, while high-frequency chemical shifts are more likely to be underestimated computationally (as experienced by Bühl et al. with the B3LYP functional).11 The 61Ni−31P J coupling constants agree well with the experiment,7 including the value measured for Ni[P(OPh)3]4 in this work (see above), and are possibly related to the electronegativity of the substituent on the phosphorus site, as suggested by Hao et al. from solution NMR experiments.7 EFG tensor information from ZORA-DFT calculations is presented in Table 4. The calculated 61Ni quadrupolar coupling constants follow the trends predicted by symmetry and ligand bulk for the tetrakis(trimethylphosphine)nickel(0) and tetrakis(trimethylphosphite)nickel(0) pair (namely, that for isostructural species, the electron-withdrawing ability of the substituents is expected to have an effect on CQ),47 with Ni(PMe3)4 predicted to have the smaller quadrupolar coupling constant (on the order of tens to hundreds of kilohertz) relative to Ni[P(OMe)3]4 (on the order of a few megahertz). On the other hand, the calculated Ni(PPh3)4 and Ni[P(OPh)3]4 CQ values do not mirror the methyl-derivative trend, which is at odds with what was observed experimentally. On the basis of the NMR spectra presented in this work, it appears that the crystal structure plays a significant role in the solid-state chemistry of these compounds in particular, which is not captured by single-molecule gas-phase computations. The discrepancies between the computational CQ and η values in this work versus that of Bühl et al.11 (Table 4), particularly for Ni(cod)2, suggest that the computational 61Ni NMR results are highly dependent on very small structural changes but are,

only at one level, they are equivalent within the second decimal place to those of Bühl et al.11). GaussView40 plots of pre- and postoptimization models highlight the differences in the lowestenergy conformations of the −(Ph3)4 and −[(OPh)3]4 pair and are provided in the Supporting Information. The significance of the computational results, apart from validating the assignment of the experimental spectrum, is that their implications correspond with Tolman’s original assertion that steric, rather than electronic, effects can determine homogeneous catalytic structure and function,44,46 demonstrating the principle for one of the compounds that he used in his original investigations into ligand cone angles and explaining the particular instability of Ni(PPh3)4. The equilibrium of Scheme 1 and its solvent dependence are readily observable in solution 31P and 13C NMR spectra (Figure 9), and the temperature dependence of the equilibrium was previously characterized by NMR studies (among other methods).45 For both nuclides, the appearance of relatively high-integration and broad peaks suggests a strong influence from rapid ligand exchange, with the proportion of signal stemming from the trigonal-planar and tetrahedral species being strongly solvent-dependent: in benzene-d6, most of the phosphorus signal in the sample stems from the broadened, tetrahedral species line shape; in acetone-d6 and acetonitrile-d3, the equilibrium appears to favor the trigonal-planar form. In the case of acetonitrile-d3, in particular, the peak areas of the free ligand and other signal are 1.0−3.0, which, while taking into account the uncertainty in the integration, is what would be expected from dissociation of Ni(PPh3)4. Benzene-d6 was selected to attempt characterization of the solution 61Ni NMR spectrum of Ni(PPh3)4, but no signal was obtained at 21.1 T after many hours of acquisition, likely because of broadening of the signal under rapid ligand exchange. At this time, it is unlikely that solution 61Ni NMR spectroscopy is possible for this compound at room temperature because of the interplay of unfavorable kinetics and thermodynamics for this species in a wide range of solvents. However, in the future, it may be possible to extract the J coupling pattern for this species by running experiments at low temperature, where it has been shown by 31P NMR in solution that the rate of exchange is slowed to manageable levels.45 Computational Details. A summary of the chemical shift and J coupling results obtained from ZORA-DFT calculations for a series of nickel compounds is provided in Table 5. As discussed in the Experimental Section, the chemical shifts were 10004


Inorganic Chemistry

■

nevertheless, useful for anticipating the order of magnitude of the interactions involved.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.inorgchem.7b01536. Table of probes used, 21.1 and 11.7 T 61Ni static NMR spectra for Ni(cod)2, temperature-dependent plot of δiso(61Ni) for Ni(cod)2 saturated in C6D6, 31P MAS NMR spectra at two fields for Ni(PPh3)4, plot of computational versus experimental δiso(61Ni), and GaussView plots for the pre- and post-DFT geometry optimization of Ni(PPh3)4 and Ni[P(OPh3)4] (PDF)

■

REFERENCES

(1) Bastow, T. J.; West, G. W. Defects and Hyperfine Interactions in Ni−Y Intermetallics (Y = Al, Ga, In, Ti) via 27Al, 47Ti, 61Ni, 69,71Ga and 115 In Nuclear Resonance. J. Phys.: Condens. Matter 2003, 15 (49), 8389−8406. (2) Mishina, S.; Iwamoto, Y.; Kohori, Y.; Ueda, K.; Kohara, T.; Jung, M.-H.; Takabatake, T. Nuclear Magnetic Resonance and Nuclear Quadrupole Resonance Studies of 61Ni in Heavy Fermion Compounds Ce7Ni3. J. Appl. Phys. 2000, 87 (9), 5134. (3) Segransan, P. J.; Chabre, Y.; Clark, W. G. Nuclear Spin Relaxation and the Knight Shift of 61Ni Metal in the Paramagnetic Solid and Liquid Phases. J. Phys. F: Met. Phys. 1978, 8 (7), 1513−1524. (4) Segransan, P. J.; Clark, W. G.; Chabre, Y.; Carter, G. C. The Knight Shift and Nuclear Spin Relaxation of 61Ni Metal in the Paramagnetic Phase. J. Phys. F: Met. Phys. 1976, 6 (5), L153−L156. (5) Drain, L. The Magnetic Moment of 61Ni. Phys. Lett. 1964, 11, 114−115. (6) Schumann, H.; Meissner, M.; Kroth, H.-J. NMR Studies on Organoelement(IVb)-Phosphines, VI [1] 61Ni NMR Spectroscopic Investigations on Organoelement(IVb) Substituted Complexes of Tetracarbonylnickel. Z. Naturforsch., B: J. Chem. Sci. 1980, 35 (5), 639−641. (7) Hao, N.; McGlinchey, M. J.; Sayer, B. G.; Schrobilgen, G. J. A 61 Ni NMR Study of Some d10 Nickel Complexes. J. Magn. Reson. 1982, 46 (1), 158−162. (8) Behringer, K. D.; Blümel, J. 61Ni NMR Spectroscopy of Di- and Tricarbonylnickel Complexes. Magn. Reson. Chem. 1995, 33 (9), 729− 733. (9) Behringer, K. D.; Blümel, J. Immobilization of Carbonylnickel Complexes: A Solid-State NMR Study. Inorg. Chem. 1996, 35 (7), 1814−1819. (10) Benn, R.; Rufińska, A. High-Resolution Metal NMR Spectroscopy of Organometallic Compounds. Angew. Chem., Int. Ed. Engl. 1986, 25 (10), 861−881. (11) Bühl, M.; Peters, D.; Herges, R. Substituent Effects on 61Ni NMR Chemical Shifts. Dalton Trans. 2009, 30, 6037−6044. (12) Benn, R.; Rufińska, A. Indirect Two-Dimensional Heteronuclear NMR Spectroscopy of Low-γ Metal Nuclei (M = 183W, 57Fe, 103Rh, 61 Ni). Magn. Reson. Chem. 1988, 26 (10), 895−902. (13) Avent, A. G.; Cloke, F. G. N.; Day, J. P.; Seddon, E. A.; Seddon, K. R.; Smedley, S. M. Tetrakis(trimethylphosphine)nickel(0). J. Organomet. Chem. 1988, 341 (1−3), 535−541. (14) Bryce, D. L.; Wasylishen, R. E. The First Chromium-53 SolidState Nuclear Magnetic Resonance Spectra of Diamagnetic chromium(0) and chromium(VI) Compounds. Phys. Chem. Chem. Phys. 2001, 3 (23), 5154−5157. (15) Ooms, K. J.; Feindel, K. W.; Terskikh, V. V.; Wasylishen, R. E. Ultrahigh-Field NMR Spectroscopy of Quadrupolar Transition Metals: 55Mn NMR of Several Solid Manganese Carbonyls. Inorg. Chem. 2006, 45 (21), 8492−8499. (16) Herzfeld, J.; Berger, A. E. Sideband Intensities in NMR Spectra of Samples Spinning at the Magic Angle. J. Chem. Phys. 1980, 73 (12), 6021−6030. (17) Still, B. M.; Kumar, P. G. A.; Aldrich-Wright, J. R.; Price, W. S. 195 Pt NMR  Theory and Application. Chem. Soc. Rev. 2007, 36 (4), 665−686. (18) Priqueler, J. R. L.; Butler, I. S.; Rochon, F. D. An Overview of 195 Pt Nuclear Magnetic Resonance Spectroscopy. Appl. Spectrosc. Rev. 2006, 41 (3), 185−226. (19) Benn, R.; Reinhardt, R.-D.; Rufińska, A. 195Pt NMR of Organoplatinum Compounds. J. Organomet. Chem. 1985, 282 (2), 291−295. (20) Benn, R.; Michael Büch, H.; Reinhardt, R.-D. Heavy Metal Spin1/2 Nuclei. Platinum-195 Relaxation Times in Phosphorus-platinum(0) Compounds and Their Dependence on the Geometry of the Complex. Magn. Reson. Chem. 1985, 23 (7), 559−564.

CONCLUSIONS We have presented, to our knowledge, the first SSNMR spectra of 61Ni for nonmetallic species, specifically for a series of diamagnetic nickel(0) complexes. The broad 61Ni line shape of Ni(cod)2 presented provides evidence that, especially with the use of high fields, valuable spectra can be acquired even for elements with as unfavorable nuclear and chemical properties as nickel. Further, we have provided an example demonstrating that SSNMR, in conjunction with computational studies, is invaluable in characterizing disordered systems, specifically in providing an explanation for the solid-state behavior of Ni(PPh3)4 and offering insight into the crystallographic nature of Ni[P(OPh)3]4. With the advent of ever higher magnetic field spectrometers, as well as ongoing strides in dynamic nuclear polarization and indirect detection methods,48,49 it may be anticipated that the quality of 61Ni SSNMR spectra will substantially improve in the future. NMR spectroscopy of various challenging nuclides continues to be an indispensable tool particularly when other techniques prove unsuitable or insufficient in explaining organometallic chemistry. We have also proposed Ni(cod)2 saturated in C6D6 as a secondary chemical shift reference for 61Ni, which should make the field of nickel NMR in particular more user-friendly in the future.

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel: +1-613-562-5800 ext. 2018. Fax: +1-613-562-5170. ORCID

David L. Bryce: 0000-0001-9989-796X Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS D.L.B. is grateful to the Natural Sciences and Engineering Research Council (NSERC) of Canada for funding. We thank Dr. Victor Terskikh for helpful discussions, the provision of 7 mm borosilicate L-shaped tubes, and assistance with experiments at 21.1 T. Access to the 21.1 T NMR spectrometer was provided by the National Ultrahigh-Field NMR Facility for Solids (Ottawa, Canada), a national research facility funded by a consortium of Canadian Universities, supported by the National Research Council of Canada and Bruker BioSpin and managed by the University of Ottawa (http://nmr900.ca). 10005


Article

Inorganic Chemistry

LNi(CO)3, LCr(CO)5, and Eta-(C5H5)Mn(CO)2L Complexes. J. Organomet. Chem. 1975, 101, 63−69. (44) Tolman, C. A.; Seidel, W. C.; Gerlach, D. H. Triarylphosphine and Ethylene Complexes of Zerovalent Nickel, Palladium, and Platinum. J. Am. Chem. Soc. 1972, 94 (8), 2669−2676. (45) Mynott, R.; Mollbach, A.; Wilke, G. A Proof of the Existence of Ni[P(C6H5)3]4 in Solution. J. Organomet. Chem. 1980, 199 (1), 107− 109. (46) Tolman, C. A. Phosphorus Ligand Exchange Equilibriums on Zerovalent Nickel. A Dominant Role for Steric Effects. J. Am. Chem. Soc. 1970, 92 (10), 2956−2965. (47) Autschbach, J.; Zheng, S.; Schurko, R. W. Analysis of Electric Field Gradient Tensors at Quadrupolar Nuclei in Common Structural Motifs. Concepts Magn. Reson., Part A 2010, 36A (2), 84−126. (48) Venkatesh, A.; Hanrahan, M. P.; Rossini, A. J. Proton Detection of MAS Solid-State NMR Spectra of Half-Integer Quadrupolar Nuclei. Solid State Nucl. Magn. Reson. 2017, DOI: 10.1016/ j.ssnmr.2017.03.005. (49) Perras, F. A.; Kobayashi, T.; Pruski, M. Magnetic Resonance Imaging of DNP Enhancements in a Rotor Spinning at the Magic Angle. J. Magn. Reson. 2016, 264, 125−130. (50) Hedberg, L.; Iijima, T.; Hedberg, K. Nickel Tetracarbonyl, Ni(CO)4. I. Molecular Structure by Gaseous Electron Diffraction. II. Refinement of Quadratic Force Field. J. Chem. Phys. 1979, 70 (7), 3224−3229. (51) Nickel, T.; Goddard, R.; Krü ger, C.; Pö rschke, K.-R. Cyclotrimerization of Ethyne on the Complex Fragment[(η1tBu2PCH2PtBu2)Ni(0)] with Formation of an η6-Benzene-Nickel(0) Complex. Angew. Chem., Int. Ed. Engl. 1994, 33 (8), 879−882. (52) Almenningen, A.; Andersen, B.; Astrup, E. E.; et al. An Electron Diffraction Investigation of Nickel Tetrakis-Trifluorophosphine, Ni(PF3)4, in the Vapour Phase. Acta Chem. Scand. 1970, 24, 1579−1584. (53) Dick, D. G.; Stephan, D. W.; Campana, C. F. The Crystal and Molecular Structure of the Coordinatively Unsaturated Ni(0) Species Ni(PPh3)3. Can. J. Chem. 1990, 68, 628−632.

(21) Dechter, J. NMR of Metal Nuclides. Part II: The Transition Metals. In Progress in Inorganic Chemistry; Lippard, S. J., Ed.; John Wiley & Sons, 1985; Vol. 33; p 393. (22) Rüegger, H. Indirect Coupling to Quadrupolar Nuclei: Observation of 1J(105Pd,31P) in One- and Two-Dimensional 31P CP/ MAS Spectra. Magn. Reson. Chem. 2004, 42 (9), 814−818. (23) Seitchik, J. A.; Gossard, A. C.; Jaccarino, V. Knight Shifts and Susceptibilities of Transition Metals: Palladium. Phys. Rev. 1964, 136 (4A), A1119−A1125. (24) Matsuda, K.; Kohori, Y.; Kohara, T. Observation of 105Pd NMR and NQR Signals in the Heavy-Fermion Superconductor UPd2Al3. Phys. Rev. B: Condens. Matter Mater. Phys. 1997, 55 (22), 15223− 15227. (25) Fedotov, M. A.; Likholobov, V. A. First Direct Observation of 105 Pd NMR in Solution. Bull. Acad. Sci. USSR, Div. Chem. Sci. 1984, 33, 1751. (26) Wagler, T. A.; Daunch, W. A.; Rinaldi, P. L.; Palmer, A. R. Solid State 33S NMR of Inorganic Sulfides. J. Magn. Reson. 2003, 161 (2), 191−197. (27) O’Dell, L. A.; Moudrakovski, I. L. Testing the Sensitivity Limits of 33S NMR: An Ultra-Wideline Study of Elemental Sulfur. J. Magn. Reson. 2010, 207 (2), 345−347. (28) Behringer, K. D.; Blümel, J. Immobilization and Chelation of Metal Complexes with Bifunctional Phosphine Ligands: A Solid-State NMR Study. Chem. Commun. 1996, 5, 653−654. (29) Tasker, S. Z.; Standley, E. A.; Jamison, T. F. Recent Advances in Homogeneous Nickel Catalysis. Nature 2014, 509, 299−309. (30) Qiu, Z. Late Transition Metal-Carboryne Complexes: Synthesis, Structure, Bonding, and Reaction with Alkenes and Alkynes; Springer, 2012. (31) Iyer, S.; Sattar, A. K. Ni[P(OPh)3]4 Catalyzed Transfer Hydrogenation Using HCOONH4. Synth. Commun. 1998, 28 (10), 1721−1725. (32) Brandsma, L.; Vasilevsky, S. F.; Verkruijsse, H. D. Nickel- and Palladium-Catalyzed Cross-Coupling Reactions with Organometallic Intermediates. Application of Transition Metal Catalysts in Organic Synthesis; Springer, 1999; p 227. (33) Lyons, J. E. Transition Metal Complexes as Catalysts for the Addition of Oxygen to Reactive Organic Substrates. In Aspects of Homogeneous Catalysis: A Series of Advances; Ugo, R., Ed.; Springer: Amsterdam, The Netherlands, 1977; Vol. 3, p 29. (34) Wender, P. A.; Smith, T. E. Bis(1,5-cyclooctadiene)nickel(0). Encyclopedia of Reagents for Organic Synthesis; John Wiley & Sons, Ltd.: Chichester, U.K., 2001. (35) Harris, R. K.; Becker, E. D.; Cabral de Menezes, S. M.; Goodfellow, R.; Granger, P. NMR Nomenclature. Nuclear Spin Properties and Conventions for Chemical Shifts (IUPAC Recommendations 2001). Pure Appl. Chem. 2001, 73 (11), 1795−1818. (36) Ammann, C.; Meier, P.; Merbach, A. E. A Simple Multinuclear NMR Thermometer. J. Magn. Reson. 1982, 46 (2), 319−321. (37) ADF2016. SCM. Theoretical Chemistry at Vrije Universiteit, Amsterdam, The Netherlands, http://www.scm.com. (38) te Velde, G.; Bickelhaupt, F. M.; Baerends, E. J.; Fonseca Guerra, C.; van Gisbergen, S. J. A.; Snijders, J. G.; Ziegler, T. Chemistry with ADF. J. Comput. Chem. 2001, 22 (9), 931−967. (39) Fonseca Guerra, C.; Snijders, J. G.; te Velde, G.; Baerends, E. J. Towards an Order-N DFT Method. Theor. Chem. Acc. 1998, 99 (6), 391−403. (40) Dennington, R.; Keith, T. A.; Millam, J. M. GaussView, version 4; Semichem Inc.: Shawnee Mission, KS, 2016. (41) Macchi, P.; Proserpio, D. M.; Sironi, A. Experimental Electron Density Studies for Investigating the Metal π-Ligand Bond: The Case of Bis(1,5-Cyclooctadiene)nickel. J. Am. Chem. Soc. 1998, 120 (7), 1447−1455. (42) Blaakmeer, E. S.; Antinucci, G.; Busico, V.; van Eck, E. R. H.; Kentgens, A. P. M. Solid-State NMR Investigations of MgCl2 Catalyst Support. J. Phys. Chem. C 2016, 120 (11), 6063−6074. (43) Bodner, G. M.; Gaul, M. M. A Carbon-13 NMR Study of Phosphine, Phosphite, Arsine, and Stibine Ligands and Their 10006


Ni Solid-State NMR Spectroscopy - ACS Publications - American

Recommend Documents