Solid-State NMR Techniques for the Structural Characterization of Cyclic Aggregates Based on Borane–Phosphane Frustrated Lewis Pairs

molecules

Review Solid-State NMR Techniques for the Structural Characterization of Cyclic Aggregates Based on Borane–Phosphane Frustrated Lewis Pairs

Robert Knitsch 1, Melanie Brinkkötter 1, Thomas Wiegand 2, Gerald Kehr 3 , Gerhard Erker 3, Michael Ryan Hansen 1 and Hellmut Eckert 1,4,* 1 Institut für Physikalische Chemie, WWU Münster, 48149 Münster, Germany; [email protected] (R.K.); [email protected] (M.B.); [email protected] (M.R.H.) 2 Laboratorium für Physikalische Chemie, ETH Zürich, 8093 Zürich, Switzerland; [email protected] 3 Organisch-Chemisches Institut, WWU Münster, 48149 Münster, Germany; [email protected] (G.K.); [email protected] (G.E.) 4 Instituto de Física de Sao Carlos, Universidad de Sao Paulo, Sao Carlos SP 13566-590, Brazil * Correspondence: [email protected]

Academic Editor: Mattias Edén Received: 20 February 2020; Accepted: 17 March 2020; Published: 19 March 2020

Abstract: Modern solid-state NMR techniques oﬀer a wide range of opportunities for the structural characterization of frustrated Lewis pairs (FLPs), their aggregates, and the products of cooperative addition reactions at their two Lewis centers. This information is extremely valuable for materials that elude structural characterization by X-ray diﬀraction because of their nanocrystalline or amorphous character, (pseudo-)polymorphism, or other types of disordering phenomena inherent in the solid state. Aside from simple chemical shift measurements using single-pulse or cross-polarization/magic-angle spinning NMR detection techniques, the availability of advanced multidimensional and double-resonance NMR methods greatly deepened the informational content of these experiments. In particular, methods quantifying the magnetic dipole–dipole interaction strengths and indirect spin–spin interactions prove useful for the measurement of intermolecular association, connectivity, assessment of FLP–ligand distributions, and the stereochemistry of adducts. The present review illustrates several important solid-state NMR methods with some insightful applications to open questions in FLP chemistry, with a particular focus on supramolecular associates.

Keywords: Frustrated Lewis pairs; aggregation; solid-state NMR; dipolar spectroscopy; internuclear distance measurement

1. Introduction During the past decade, borane–phosphane frustrated Lewis pairs (FLPs) attracted great interest as metal-free systems for a large variety of catalytic and stoichiometric chemical transformations [1–17]. Their Lewis centers can be incorporated in the same molecule (intramolecular FLPs) or in two diﬀerent entities (intermolecular FLPs). In these molecules, both Lewis centers are shielded by sterically demanding groups, which prevent quenching, i.e., the formation of a covalent bond between them. This so-called frustration phenomenon results in a remarkable reactivity rarely encountered in metal-free organic molecules. The most prominent example of such a reaction is the ability of borane–phosphane (B/P) FLPs to split dihydrogen and to transfer the resulting proton/hydride pair to other organic substrates such as unsaturated carbon bonds [1–7]. In addition to their ability of splitting dihydrogen, B/P FLPs can give rise to various types of adducts with a large number of small molecules, leading to interesting heterocycles and complex organic structures, which might be valuable building

Molecules 2020, 25, 1400; doi:10.3390/molecules25061400 www.mdpi.com/journal/molecules Molecules 2019, 24, x FOR PEER REVIEW 2 of 40

organic substrates such as unsaturated carbon bonds [1–7]. In addition to their ability of splitting

Moleculesdihydrogen,2020, 25 B/P, 1400 FLPs can give rise to various types of adducts with a large number of small2 of 39 molecules, leading to interesting heterocycles and complex organic structures, which might be valuable building blocks for novel organoborane materials. These small molecules include CO2, CO, blocksSO2, NO, for alkenes, novel organoborane alkynes, carbonyls, materials. isonitriles, Thesesmall and many molecules more, include as summarize CO2, CO,d in SO Figure2, NO, 1 [1,7–17]. alkenes, alkynes,Beginning carbonyls, in 2010, isonitriles,we developed and a many comprehensive more, as summarized research program in Figure exploring1[ 1,7–17 the]. Beginning relationship in 2010,between we developedstructure aand comprehensive chemical activity research of program these systems, exploring using the relationship advanced betweensolid-state structure NMR andtechniques chemical [18–20]. activity These of these techniques systems, were using not advanced only found solid-state to assist in NMR the structural techniques characterization [18–20]. These techniquesof disordered were solids not only that found elude to single-crystal assist in the structural diffraction, characterization but to provide of disordered suitable anisotropic solids that eludeobservables single-crystal (magnetic diff raction,shielding, but direct to provide and suitableindirect anisotropic spin–spin observablescoupling, and (magnetic nuclear shielding, electric directquadrupolar and indirect coupling spin–spin tensors) coupling, that are and able nuclear to characterize electric quadrupolar the centers coupling individually, tensors) including that are able the toextent characterize of intramolecular the centers electronic individually, interactions including between the extent them of intramolecular[20–22]. As such, electronic NMR spectroscopy interactions betweenwas helpful them to [20predict–22]. Asthe such,degree NMR of frustration spectroscopy of such was systems, helpful which to predict rela thetes degreeto their of reactivity. frustration ofIn suchaddition, systems, NMR which methods relates proved to their reactivity.to be instrumental In addition, in NMR characterizing methods proved reaction to bepathways instrumental and inintermediates characterizing associated reaction with pathways FLP chemistry, and intermediates as well as associatedfor the stereochemical with FLP chemistry, analysis of as the well final as forproducts the stereochemical obtained [12,21–25]. analysis ofAs the discussed final products in previous obtained reviews [12,21– 25[18,19],]. As discussed NMR spectroscopic in previous reviewsparameters [18, 19allow], NMR a distinction spectroscopic between parameters classical allow and a distinctioncooperative between adducts, classical the identification and cooperative and adducts,quantification the identification of different andpossible quantification isomers and of diepimers,fferent possibleand the isomersdescription and of epimers, intermolecular and the descriptionaggregation of and intermolecular oligomerization aggregation effects. and oligomerization effects.

Figure 1. 1. SchematicSchematic overview overview for various for various chemical chemical reactions reactions performed performed with intramolecular with intramolecular borane– borane–phosphanephosphane (B/P) frustrated (B/P) frustrated Lewis pairs Lewis (FLPs). pairs (FLPs).

The formationformation of of supramolecular supramolecular macrocyclic macrocyclic structures structures by FLPs by andFLPs their and small-molecule their small-molecule adducts wasadducts developed was developed further duringfurther theduring past the ﬁve past years, five and years, it forms and it the forms focus the of focus the present of the present review. Suchreview. supramolecular Such supramolecular assemblies assemblies are not onlyare not interesting only interesting in their ownin their right own from right a sheer from esthetics a sheer viewpoint,esthetics viewpoint, but also because but also they because represent they smallrepres spin-clusterent small spin-cluster systems, allowing systems, advanced allowing solid-stateadvanced NMRsolid-state methodologies NMR methodologies to be tested, to validated, be tested, and validat optimizeded, and with optimized the help with of theoretical the help of calculations theoretical andcalculations simulations. and Insimulations. the present In review, the present we illustrate review, how we advanced illustrate solid-state how advanced NMR methodologysolid-state NMR can bemethodology used to obtain can essential be used information to obtain onessential the structural information organization on the ofstructural such complex organization materials, of which such provide important insights into their reactivity and self-assembling features. The use of solid-state NMR spectroscopy is crucial for these purposes, as the supramolecular arrangements may fall apart Molecules 2020, 25, 1400 3 of 39 into their monomeric constituents upon dissolution. Furthermore, the anisotropic interactions, like dipolar and quadrupolar couplings, which are most intimately linked to structural information, are not accessible in the liquid state. Here, we focus on measurements of crystallographically well-characterized materials, thereby laying the foundations for future applications to FLP-based systems in poly-, nano-, or disordered crystalline and amorphous states of matter.

2. Solid-State NMR—Basics and Methodology

2.1. Fundamental Principles The basic principles of nuclear magnetic resonance (NMR) were summarized in numerous excellent books and review articles devoted to applications to various ﬁelds in solid-state chemistry [26–30]. The method is based on nuclear spin motion, represented by an angular momentum operator J.

J = I} (1) where } is the reduced Planck’s constant, and I is a dimensionless operator, satisfying angular momentum commutation rules and quantization conditions.

I2 = I(I + 1) (2)

Iz = m ε I, I 1, ... , I . (3) | | { − − } Thus, the nuclear spin angular momentum is orientationally quantized into 2I + 1 degenerate states. Here, I denotes the nuclear spin quantum number, which is also nucleus-specific and adopts one of the half-integer values 1/2, 3/2, 5/2, 7/2, or 9/2 for stable nuclei (with the exception of a few nuclei with integer spin). Nuclear spin angular momentum entails nuclear magnetism, according to the following relationship: µ = γI} (4) where γ, the gyromagnetic ratio, is a nuclear specific constant. In NMR, these nuclear magnetic moments µ can be detected by applying an external magnetic field, represented by the magnetic flux density B0. The resulting Zeeman interaction lifts the energetic equivalence of orientational states and causes a splitting into 2I + 1 individual levels with energies

Em = mγ}B0 (5) − where m denotes the orientational quantum numbers, which, in the classical picture, correspond to spin orientations that are partially aligned or counter-aligned with the direction of B0. By application of electromagnetic waves satisfying the Bohr condition ∆E = }ω, allowed transitions between adjacent states (∆m = 1) can be stimulated and observed spectroscopically. The interaction takes place between ± the nuclear magnetic dipole moments and the oscillating magnetic ﬁeld of the applied electromagnetic wave. The angular resonance frequency is given by

ω = γB0 (6) which is equal to 2πυP = γB0, where υP is the Larmor frequency with which the nuclei precess in the applied magnetic field. Since the values of γ differ greatly for different kinds of nuclei, NMR spectroscopy is intrinsically element-selective as, at a given value of B0, different nuclear isotopes have different resonance and precession frequency ranges. With typical values of applied magnetic flux densities between 4.65 and 23.6 T used in NMR, the frequencies lie in the radio-wave region (10–1000 MHz), depending on the value of γ. The precession frequencies are measured using pulsed excitation, recording the resulting free induction decay (FID). Fourier transformation then produces the frequency domain signal. Molecules 2019, 24, x FOR PEER REVIEW 4 of 40

While nuclear precession frequencies measured in NMR are usually dominated by the Zeeman interaction, they are additionally influenced by a number of internal interactions, whose parameters reflect the details of the local structural environment. The electronic environments around the nuclei produce magnetic shielding effects that modify the local fields experienced by the nuclei. These electronic environments also produce electric field gradients (EFGs), which can interact with the nuclear electric quadrupole moments of spin >½ nuclei. In addition, the electronic environment can promote internuclear spin–spin coupling by a spin-polarization mechanism (through-bond coupling). Finally, even in the absence of bond connectivity, the local magnetic fields are altered by direct Molecules 2020, 25, 1400 4 of 39 (through-space) dipole-dipole coupling, whose strength is proportional to the inverse cubed internuclear distance. The strengths and spin dynamics of the latter two interactions further depend on whether 2.2.the Internal interacting Nuclear nuclei Interactions are of the same or of different kinds (homo- versus heteronuclear couplings). Figure 2 summarizes the internal interactions, their most important NMR parameters, and their While nuclear precession frequencies measured in NMR are usually dominated by the Zeeman relevance for structure elucidation. They refer to the common case of diamagnetic solids; additional interaction, they are additionally influenced by a number of internal interactions, whose parameters interactions are present in paramagnetic materials [31]. reflect the details of the local structural environment. The electronic environments around the In the solid state, all internal nuclear spin interactions are anisotropic, which means that their nuclei produce magnetic shielding effects that modify the local fields experienced by the nuclei. These effects on the nuclear precession frequency depend on molecular orientation in the magnetic field. electronic environments also produce electric field gradients (EFGs), which can interact with the Thus, the NMR spectra of polycrystalline samples are significantly broadened, owing to the statistical nuclear electric quadrupole moments of spin > 1 nuclei. In addition, the electronic environment can orientation distribution of the crystallites. Furthermore,2 the four internal interactions (Figure 2) are promote internuclear spin–spin coupling by a spin-polarization mechanism (through-bond coupling). usually superimposed, which makes the resulting NMR spectra difficult to analyze. However, as Finally, even in the absence of bond connectivity, the local magnetic fields are altered by direct detailed further below, this superposition can be disentangled by special selective averaging (through-space) dipole-dipole coupling, whose strength is proportional to the inverse cubed internuclear experiments. This disentanglement is usually essential for linking the spectroscopic details to distance. The strengths and spin dynamics of the latter two interactions further depend on whether the structural information. The following sections introduce the features of the four internal interaction interacting nuclei are of the same or of different kinds (homo- versus heteronuclear couplings). Figure2 mechanisms and their link to structural information. Alongside, we describe the experimental summarizes the internal interactions, their most important NMR parameters, and their relevance for strategies that are most commonly used in FLP chemistry for extracting such interactions in a structure elucidation. They refer to the common case of diamagnetic solids; additional interactions are selective fashion. present in paramagnetic materials [31].

Figure 2. Anisotropic interactions inﬂuencing the appearance of solid-state NMR spectra for diamagnetic solids, parameters extractable from the NMR spectra describing these interactions, and the structural information they are connected to.

In the solid state, all internal nuclear spin interactions are anisotropic, which means that their effects on the nuclear precession frequency depend on molecular orientation in the magnetic field. Thus, the NMR spectra of polycrystalline samples are significantly broadened, owing to the statistical orientation distribution of the crystallites. Furthermore, the four internal interactions (Figure2) are usually superimposed, which makes the resulting NMR spectra difficult to analyze. However, as detailed further below, this superposition can be disentangled by special selective averaging experiments. This disentanglement is usually essential for linking the spectroscopic details to structural information. Molecules 2020, 25, 1400 5 of 39

The following sections introduce the features of the four internal interaction mechanisms and their link to structural information. Alongside, we describe the experimental strategies that are most commonly used in FLP chemistry for extracting such interactions in a selective fashion.

2.2.1. Magnetic Shielding Eﬀects

Basic Principles: The externally applied magnetic field B0 induces magnetic polarization effects in the core and valence shell electronic environment of the nuclei, influencing their nuclear precession frequencies. This anisotropic effect, called magnetic shielding, is described by a second-rank tensor, which can be parametrized in terms of the following three parameters: σiso = 1/3(σ33 + σ22+ σ11), which corresponds to the isotropic average value; ∆σ = σ 1/2(σ + σ ), its anisotropy and ησ = (σ 33 − 11 22 22 − σ )/(σ σ ), its asymmetry parameter, that is, the deviation of the tensor from axial symmetry [32]. 11 33 − iso In the simplest case of an axially symmetric magnetic shielding anisotropy, ωP is given by 1 2 ωP(θ) = 2πν (θ) = γB 1 σ ∆σ 3cos θ 1 , (7) P 0 − iso − 3 − where the angle θ specifies the orientation for the axis of the magnetic shielding main component σ33 relative to the magnetic field direction. Note that, for θ = 54.7 , the term 3cos2θ 1 is zero, and only ◦ − the isotropic value is measured. Experimental Determination: The anisotropy of ωP expressed by Equation (7) can also be removed if the time average of the term 3cos2θ 1 can be brought to zero. Specifically, this is achieved by the − magic-angle sample spinning (MAS) technique, where the NMR spectrum is measured while rapidly rotating the sample about an axis inclined by an angle of βm = 54.7◦ relative to the magnetic field direction [33]. This manipulation creates the same average orientation < θ> of 54.7◦ for all molecules or structural fragments (irrespective of their original orientation), if rotation is sufficiently rapid, leaving the same isotropic average of υP to be measured for all spins in the same chemical environment. The result is a significant line-narrowing effect allowing the identification of distinct local environments on the basis of their individual isotropic magnetic shielding constants σiso. Magnetic shielding values can currently be calculated with high accuracy using suitable quantum chemical methods [34]. Experimentally accessible values, however, are not absolute shielding values (which would require comparisons with bare nuclei) but rather chemical shifts measured relative to the precession frequency of a reference compound. υsample υre f δiso = − . (8) υre f

If the rotation frequency νR does not fulfill the condition 2πνR >> γB0 ∆σ, the narrowed NMR signal is accompanied by spinning sidebands, which are spaced from the central signal at integer multiples of νR (see Figure3). When quantifying relative ratios of spectroscopically resolved species, the areas of these spinning sidebands must also be taken into account. Furthermore, in the slow-spinning limit (2πνR << γB0 ∆σ), accurate values of ∆σ and ησ can be extracted from the intensity profiles, providing additional information regarding the local symmetry of the nuclear species observed [35]. On the other hand, the spinning frequency required for obtaining simple single-peak-per-site spectra increases linearly with B0, as the spectral dispersion (in Hz) caused by the chemical shift anisotropy is proportional to the applied magnetic field strength. Currently, such experiments are performed at MAS frequencies up to ~110 kHz achieved with commercially available rotors of 0.7 mm diameter, although even higher MAS frequencies using even smaller-sized rotors were reported recently [36,37]. Molecules 20202019, 2524, 1400x FOR PEER REVIEW 66 of of 40 39

Figure 3. Numerical simulations demonstrating the effect effect of magic-angle spinning (MAS) upon the solid-state NMR line shape of a spin-1/2 spin-1/2 nucleus withwith dominant line broadening due to the magnetic shielding anisotropy. Spinning sidebandssidebands areare locatedlocated atat integerinteger multiplesmultiples ofof thethe rotorrotor frequency.frequency. 2.2.2. Nuclear Electric Quadrupole Interactions 2.2.2. Nuclear Electric Quadrupole Interactions Basic Principles [38]: For nuclei with a spin quantum number I > 1 , the charge distribution Basic Principles [38]: For nuclei with a spin quantum number I > ½, the2 charge distribution is not is not spherically symmetric. This asymmetry can be described by an electric quadrupole moment spherically symmetric. This asymmetry can be described by an electric quadrupole moment superimposed upon a sphere containing the nuclear charge. Classical physics predicts an interaction superimposed upon a sphere containing the nuclear charge. Classical physics predicts an interaction of these moments with inhomogeneous electric fields, i.e., electric field gradients V (EFGs), present of these moments with inhomogeneous electric fields, i.e., electric field gradients 𝑉 ij (EFGs), present at the nuclear sites. The latter are generated by the asymmetric charge distributions associated with at the nuclear sites. The latter are generated by the asymmetric charge distributions associated with the electrons of the atom bearing the nucleus, as well as atomic coordination and chemical bonding the electrons of the atom bearing the nucleus, as well as atomic coordination and chemical bonding effects. The EFG is a symmetric second-rank tensor, which can be diagonalized in a molecular axis effects. The EFG is a symmetric second-rank tensor, which can be diagonalized in a molecular axis system. The sum of the diagonal elements Vxx + Vyy + Vzz vanishes (Laplace equation) such that the system. The sum of the diagonal elements 𝑉 +𝑉 +𝑉 vanishes (Laplace equation) such that the interaction can be described in terms of two parameters, the nuclear electric quadrupolar coupling interaction can be described in terms of two parameters, the nuclear electric quadrupolar coupling constant (in frequency units), constant (in frequency units), eQVzz CQ = , (9) 𝑒𝑄𝑉 h 𝐶 = , (9) and the asymmetry parameter, ℎ Vxx Vyy and the asymmetry parameter, ηQ = − , (10) Vzz where 0 ηQ 1 describes the deviation of the𝑉 EFG−𝑉 tensor from cylindrical (axial) symmetry. Figure4 ≤ ≤ 𝜂 = , (10) illustrates typical electronic environments in molecular𝑉 inorganic chemistry and their associated

EFGwhere properties. 0 ≤ 𝜂≤ 1 describes the deviation of the EFG tensor from cylindrical (axial) symmetry. Figure 4 illustratesExperimental typical Determination: electronic environments The quadrupole in molecular interaction inorganic competes chemistry with the and Zeeman their interactionassociated forEFG quantized properties. spin alignment. For the discussion of NMR spectra, the case of a dominant Zeeman interactionExperimental is important Determination: where simple The perturbation quadrupole theory interaction can be competes applied. Itwith produces the Zeeman anisotropic interaction shifts infor the quantized Zeeman spin energy alignment. levels, which For the are discussion proportional of NMR to the spectra, square, them2, case of the of Zeeman a dominant orientational Zeeman quantuminteraction number. is important As a consequence,where simple inperturbation the case of verytheory weak can quadrupolebe applied. interactions,It produces anisotropic the central shiftsm = 1 /in2 the mZeeman= 1/2 energytransition levels, is una whichﬀected are and proportional can be analyzed to the like square, a signal fromm2, of spin-1 the /Zeeman2 nuclei. ↔ − Inorientational contrast, the quantum other allowed number.∆m As= a 1consequence, transitions (the in the so-called case ofsatellite very weak transitions quadrupole) are anisotropically interactions, ± broadenedthe central and𝑚 may = 1/2 be (in the↔ case𝑚 =−1/2of transition stronger quadrupoleis unaffected couplings) and can shiftedbe analyzed so strongly like a o ﬀsignal-resonance from thatspin-1/2 they nuclei. are only In incompletelycontrast, the other excited allowed by the Δ radiofrequencym = ±1 transitions pulses. (the so-called Stronger quadrupolesatellite transitions couplings) are willanisotropically also lead to broadened more complicated and may orientationbe (in the case dependences of stronger for quadrupole the central transition,couplings) dueshifted to theso strongly off-resonance that they are only incompletely excited by the radiofrequency pulses. Stronger

Molecules 2019, 24, x FOR PEER REVIEW 7 of 40 Molecules 2020, 25, 1400 7 of 39 quadrupole couplings will also lead to more complicated orientation dependences for the central transition, due to the presence of higher-rank tensors in the quadrupolar Hamiltonian, producing presence of higher-rank tensors in the quadrupolar Hamiltonian, producing line-broadening effects line-broadening effects that scale with the inverse of the strength of the applied magnetic field (second- that scale with the inverse of the strength of the applied magnetic field (second-order effects), and order effects), and that can be only partially averaged out by magic angle spinning [38]. For crystalline that can be only partially averaged out by magic angle spinning [38]. For crystalline compounds compounds with well-defined lattice positions or bonding geometries, often highly characteristic line with well-defined lattice positions or bonding geometries, often highly characteristic line shapes shapes can be observed, from which 𝐶 and 𝜂 can be extracted via line-shape simulation (see can be observed, from which C and η can be extracted via line-shape simulation (see Figure4), Figure 4), using standard programQ packagesQ such as DMFit [39], SIMPSON [40], or QUEST [41]. In using standard program packages such as DMFit [39], SIMPSON [40], or QUEST [41]. In addition, addition, first-principles calculations of EFG tensor values are possible using standard theoretical first-principles calculations of EFG tensor values are possible using standard theoretical programs such programs such as GAUSSIAN [42], TURBOMOLE [43], WIEN2k [44], or CASTEP [45]. as GAUSSIAN [42], TURBOMOLE [43], WIEN2k [44], or CASTEP [45].

Figure 4. (a) Typical bonding geometries realized in molecular inorganic chemistry and characterization Figure 4. (a) Typical bonding geometries realized in molecular inorganic chemistry and of their local symmetries. Geometries underlined in orange produce isotropic magnetic shielding characterization of their local symmetries. Geometries underlined in orange produce isotropic tensors (∆σ = 0) and zero electric field gradients (Vzz = Vxx = Vyy = ησ = 0; hence, C = η = 0) at magnetic shielding tensors (Δσ = 0) and zero electric field gradients (𝑉 =𝑉 =𝑉 =𝜂Q =Q0; hence, the central nucleus under consideration. Geometries underlined in yellow produce axially symmetric 𝐶 =𝜂 =0) at the central nucleus under consideration. Geometries underlined in yellow produce magnetic shielding and EFG tensors (ησ and ηQ = 0). Geometries not underlined produce asymmetric axially symmetric magnetic shielding and EFG tensors (𝜂 and 𝜂 = 0). Geometries not underlined magnetic shielding and electric field gradient tensors (0 (ησ, ηQ) 1). (b) Typical MAS NMR line produce asymmetric magnetic shielding and electric field≤ gradient≤ tensors (0 ≤ (𝜂 ,𝜂 ) ≤ 1). (b) shapes predicted by second-order perturbation theory for the m = 1 < > m = 1 central transitions of Typical MAS NMR line shapes predicted by second-order perturbation2 − theory− 2 for the m = ½ <−> m = half-integer spin nuclei due to quadrupolar interaction with EFGs of different ηQ values. −½ central transitions of half-integer spin nuclei due to quadrupolar interaction with EFGs of different An𝜂 improvedvalues. resolution for quadrupolar nuclei affected by such anisotropic line broadening is available using the multi-quantum spectroscopic (MQMAS) method [46–48]. This technique correlatesAn improved the evolution resolution of an form quadrupolarm coherence nuclei with affected the 1by/2 such anisotropic1/2 coherence line broadening (the central is | i ↔| − i | i ↔ | − i transition)available using in a two-dimensional the multi-quantum NMR spectrosco experiment.pic While(MQMAS) the excitation method of[46–48]. coherence This orders technique>1 is symmetry-forbidden,correlates the evolution this of selection an |𝑚⟩ rule↔ |−𝑚 is relaxed⟩ coherence owing with to the the substantial |½⟩ ↔ |−½ quadrupolar⟩ coherence perturbation (the central oftransition) the Zeeman in a state two-dimensional functions, such NMR that theseexperiment. coherences While can the be excitation excited in of this coherence case. Essentially, orders >1 the is designsymmetry-forbidden, of the method isthis to reverseselection the rule anisotropic is relaxed timeowing evolution to the substantial of the spin quadrupolar system during perturbation the time periodof the Zeemant1 during state the detectionfunctions, period such thatt2, leading these cohe to anrences isotropic can spectrumbe excited after in this Fourier case. transformationEssentially, the

Molecules 2019, 24, x FOR PEER REVIEW 8 of 40 design of the method is to reverse the anisotropic time evolution of the spin system during the time period t1 during the detection period t2, leading to an isotropic spectrum after Fourier transformation Molecules 2020, 25, 1400 8 of 39 with respect to the evolution time variable t1. In the spectroscopic characterization of FLPs, the most important application concerns the 11B nuclei, having spin 3/2, whose strong quadrupolar interaction results in broadened line shapes of the type shown in Figure 4b. In samples containing multiple boron with respect to the evolution time variable t1. In the spectroscopic characterization of FLPs, the most species,important this application produces concernsstrongly theoverlapping11B nuclei, MAS having NMR spin 3signals/2, whose which strong make quadrupolar the deconvolution interaction |3/2⟩ ↔ |− analysisresults indifficult broadened at times. line shapesHere, the of thetriple-quantum type shown in(TQ) Figure experiment4b. In samples involving containing the multiple 3/2⟩ borontransition species, thiscan producesproduce significant strongly overlapping improvement MAS in NMR the resolution. signals which Figure make 5 theshows deconvolution the most commonanalysis pulse difficult sequence at times. in Here,use; the the first triple-quantum pulse P1 creates (TQ) triple experiment quantum involving coherence the which3/2 is allowed3/2 | i ↔ | − i totransition evolve for can the produce time period significant t1, after improvement which it is converted in the resolution. back to Figure zero-quantum5 shows thecoherence most common by the secondpulse sequence pulse P2. in Detection use; the first follows pulse after P1 creates converting triple the quantum zero-quantum coherence to which single-quantum is allowed tocoherence, evolve for i.e., magnetization. For both pulses P1 and P2, high-radiofrequency amplitudes are required to ensure the time period t1, after which it is converted back to zero-quantum coherence by the second pulse P2. efficientDetection excitation follows afterof the converting triple quantum the zero-quantum coherence, whereas to single-quantum the detection coherence, pulse is i.e.,chosen magnetization. to be soft, toFor be both limited pulses to P1singlequantum and P2, high-radiofrequency coherence excitation amplitudes (creation are required of transverse to ensure magnetization efficient excitation from coherenceof the triple level quantum zero). For coherence, appropriate whereas coherence the detection selection, pulse special is phase chosen cycling to be soft,schemes to be have limited to be to used.singlequantum coherence excitation (creation of transverse magnetization from coherence level zero).

For appropriate coherence selection, special phase cycling schemes have to be used.

Figure 5. Pulse sequence and coherence path diagram selected via the phase cycling scheme of a Figuretriple-quantum 5. Pulse sequence MAS-NMR and experiment. coherence Forpath spin diagram 3/2 nuclei, selected coherence via the orders phase ranging cycling fromscheme+3 toof a3 − triple-quantumare possible. MAS-NMR experiment. For spin 3/2 nuclei, coherence orders ranging from +3 to −3 are possible. Figure6 shows a typical application result of an FLP octamer discussed in more detail in Section 3.6. ThisFigure compound 6 shows features a typical two application crystallographically result of an distinct FLP octamer boron discussed sites B1 and in more B2, whose detail line in Section shapes 3.6.are This strongly compound overlapping features in thetwo regular crystallographically MAS NMR spectra. distinct Double boron Fouriersites B1 transformationand B2, whose of line the shapestwo-dimensional are strongly (2D) overlapping dataset acquired in the regular with thisMAS pulse NMR sequence spectra. Double results inFourier a two-dimensional transformation map of thein whichtwo-dimensional the regular (2D) MAS dataset NMR spectrumacquired with (plotted this inpulse Figure sequence6 along results the horizontal in a two-dimensional dimension) is mapcorrelated in which with the an regular isotropic MAS spectrum NMR spectrum (plotted (plott alonged thein Figure vertical 6 along dimension). the horizontal Note thedimension) excellent isresolution correlated in with the isotropican isotropic dimension, spectrum from (plotted which along the individual the vertical MAS dimension). NMR line Note shapes the can excellent also be cs resolutionextracted byin the simulating isotropic the dimension, corresponding from F1which cross-sections, the individual yielding MAS the NMR parameters line shapesδ , canCQ ,alsoand beηQ iso extracted(see Figure by6 simulating) for both boron the corresponding sites. F1 cross-sections, yielding the parameters 𝛿,𝐶, and 𝜂 (see Figure 6) for both boron sites.

Molecules 2019, 24, x FOR PEER REVIEW 9 of 40

Molecules 2020, 25, 1400 9 of 39

Figure 6. (a) Typical TQ MAS spectrum of an FLP octamer featuring two distinct boron sites B1 and FigureB2. While 6. (a) the Typical spectrum TQ MAS in the spectrum horizontal of dimensionan FLP octamer is equivalent featuring to two the distinct regular MASboron spectrum, sites B1 and the B2.spectrum While the in thespectrum vertical in (F1) the dimensionhorizontal dimensio is purelyn isotropic, is equivalent yielding to the well-separated regular MAS symmetricspectrum, the line spectrumshapes. ( bin) Correspondingthe vertical (F1) F2 dimension sections representing is purely isotropic, resolved MASyielding spectra well-separated of the individual symmetric boron line sites shapes.including (b) lineCorresponding shape simulations F2 sections (see Section representing 3.6). resolved MAS spectra of the individual boron sites including line shape simulations (see Section 3.6). 2.2.3. Indirect Spin–Spin Coupling (“J-coupling”) 2.2.3. BasicIndirect Principles: Spin–Spin The Coupling electronic (“J-coupling”) environment of the nuclei does not only alter nuclear precession frequenciesBasic Principles: via the magnetic The electronic shielding environment or electric of quadrupolethe nuclei does coupling not only interactions, alter nuclear but precession it can also frequenciesmediate internuclear via the magnetic spin–spin shielding interactions. or electric In essence, quadrupole the spin coupling orientation interactions, of an individual but it can nucleus also mediate(I) induces internuclear a slight spin spin–spin polarization interactions. of the bonding In essence, electrons the spin linking orientation this nucleus of an withindividual another nucleus one (S), (I)altering induces the a resonanceslight spin frequencypolarization of theof the latter. bonding The e ffelectronsect is anisotropic linking this and, nucleus thus, requires with another a tensorial one (S),description altering [the49,50 resonance]. Under (sufrequencyfficiently fast)of the MAS latter. conditions, The effect only is the anisotropic isotropic componentand, thus, isrequires measured, a tensorialresulting description in a characteristic [49,50]. peak Under splitting, (sufficiently from which fast) MAS the number conditions, of bonded only the neighbors isotropic can component be inferred. 1 isIn measured, the simplest resulting case of in a a pair characteristic of two nuclei peak of sp spinlitting,2 , the from resonances which the of number both the of I andbonded the S neighbors nuclei are canthen be split inferred. into doublets,In the simplest each reflectingcase of a pair the of two two distinct nuclei possibleof spin ½, orientational the resonances states of both of the the bonded I and thenuclei. S nuclei In general,are then thesplit coupling into doublets, of an observedeach reflecting nucleus the S two to n distinctj equivalent possible nuclei orientational of spin quantum states ofnumber the bondedIj leads nuclei. to a In peak general, multiplicity the coupling given of by anΠ observedj (2njIj + nucleus1). In this S to expression, nj equivalent both nuclei homo- of spin and quantumheteronuclear number contributions Ij leads to a must peak bemultiplicity considered, given where by theΠj (2 latternjIj + can1). In be this removed expression, by radiofrequency both homo- anddecoupling heteronuclear schemes contributions (see Section 3.3 must). The be magnitude considered, of the where splitting the (in latter Hz) iscan field-independent be removed andby radiofrequencygiven by the spin decoupling–spin coupling schemes constant (see, SectionJiso. Using 3.3). this The symbol, magnitude the indirect of the splitting spin–spin (in interactionHz) is field- is independentcommonly called and givenJ-coupling by theto spin differentiate–spin coupling it from constant the through-space, Jiso. Using this dipolar symbol, coupling the indirect (seebelow). spin– spinIf quadrupolar interaction Iis nuclei commonly exposed called to strong J-coupling EFGs to are differentiate involved, the it splittingsfrom the becomethrough-space asymmetric dipolar and couplingunevenly (see spaced below). (see If Figure quadrupolar7). The simulationI nuclei exposed must thento strong include EFGs an additionalare involved, dipolar the splittings coupling becomeparameter asymmetricd, which dependsand unevenly on the spaced quadrupolar (see Figu couplingre 7). strengthThe simulation [51,52]. Themust principal then include structural an additionalinformation dipolar obtained coupling from theparameter J-coupling d, which is evidence depends for through-bondon the quadrupolar connectivity coupling (although, strength in [51,52].some special The principal cases, no-bond structural indirect information spin–spin obta couplingsined from were the also J-coupling observed is [ 53evidence,54]). The for value through- of Jiso bonddepends connectivity not only on(although, the sizes in of some the nuclear special magnetic cases, no-bond moments, indirect but also, spin–spin in particular, couplings on the were electron also observeddistribution [53,54]). in the The chemical value bondof Jiso (bonddepends covalency). not only Foron the intramolecular sizes of the borane–phosphanenuclear magnetic moments, FLPs, the 31 11 butP– also,B J-tensorsin particular, can be on calculated the electron with gooddistribu accuracytion in using the standardchemical densitybond (bond functional covalency). theory (DFT) For intramolecularmethods [20]. borane–phosphane FLPs, the 31P–11B J-tensors can be calculated with good accuracy using standard density functional theory (DFT) methods [20].

Molecules 2020, 25, 1400 10 of 39 Molecules 2019, 24, x FOR PEER REVIEW 10 of 40

FigureFigure 7. ((aa)) Exemplary Exemplary 3131PP MAS MAS NMR NMR spectrum spectrum of of a a B/P B/P FLP with isotropicisotropic 3131P J-couplings to 11BB (green)(green) and and 1010BB (red) (red) of of 33 33 and and 11 11Hz, Hz, respectively, respectively, and and residual residual dipolar dipolar couplings couplings of 8 ofand 8 and16 Hz. 16 ( Hz.b) 1 Simulated(b) Simulated MAS MAS NMR NMR spectra spectra demonstrating demonstrating the thegeneral general shape shape for foran anS =S ½= 2nucleusnucleus coupled coupled to to a a quadrupolarquadrupolar II == 3/23/2 nucleus nucleus. . Bottom spectrum: no quadrupolar interaction.interaction. Middle spectrum: with with quadrupolarquadrupolar interaction, leadingleading to to an an additional additional dipolar dipolar splitting splitting parameter parameterd arising d arising from from cross-terms cross- termsbetween between the quadrupolar the quadrupolar interaction interaction and the and heteronuclear the heteronuclear dipolar interaction.dipolar interaction. The dashed The blue dashed lines blueon the lines right on handthe right side, hand each side, located each at located a peak at maximum a peak maximum of the bottom of the or bottom middle or curve, middle show curve, the showrelative the shift relative of two shift resonances of two resona by thences residual by the dipolarresidual coupling dipolar couplingd, which d is, negativewhich is fornegative coupling for

couplingof S to I withof S mtoI =I with1/2 and,mI = ±1/2 vice and, versa, vice positive versa, for positivemI = 3for/2. m TheI = ±3/2. overall The change overall of change the resonance of the ± ± resonancefrequency ∆frequencyνm caused Δν bym isotropic causedJ -couplingby isotropic and residualJ-coupling dipolar and coupling residuald isdipolar exemplarily coupling shown d foris

exemplarilyresonances originating shown for fromresonances couplings origin of Sating to I withfromm couplingsI = 1/2 and of mS Ito= I + with3/2. Top mI = spectrum: −1/2 and simulatedmI = +3/2. − Topproﬁle spectrum: with additional simulated line profile broadening with additional to create a line line broadening shape similar to tocreate theexperimental a line shape similar curve (shown to the experimentalin green in (a )).curve (shown in green in (a)).

ExperimentalExperimental Determination: Determination: Frequently, Frequently, J-splitti J-splittingsngs such such as as those those evident evident in in Figure Figure 7 are notnot experimentallyexperimentally observable observable under under MAS MAS conditions conditions because because the the line line shapes shapes are are affected affected by by other other broadeningbroadening mechanisms.mechanisms. In thisIn case,this two-dimensionalcase, two-dimensional spectroscopic spectroscopi approachesc approaches prove instrumental. prove instrumental.The simplest methodThe simplest is the method J-based is heteronuclear the J-based heteronuclear resolved experiment resolved illustrated experiment in illustrated Figure8[ 55 in]. FigureThe method 8 [55]. exploitsThe method a time-reversal exploits a strategytime-reversal using strategy spin inversion using spin with inversion 180◦ pulses; with chemical 180° pulses; shift evolution during an initial time period t /2 is canceled by chemical shift evolution during a second chemical shift evolution during an initial1 time period t1/2 is canceled by chemical shift evolution time period t /2 following a π pulse; all the while, evolution under J-coupling continues unaffected during a second1 time period t1/2 following a π pulse; all the while, evolution under J-coupling continuesby the latter. unaffected In the heteronuclear by the latter. In (I–S) theversion, heteronuclear both types (I–S) ofversion, nuclear both spins types need of tonuclear be inverted spins need by π topulses be inverted to ensure by continuedπ pulses to evolution. ensure continued This evolution evolution. is stopped This evolution by the subsequent is stopped byπ/2 the pulses subsequent on both πchannels/2 pulses (producing on both channels z-filtering), (producing while thez-filtering), following whileπ/2 pulsesthe following create transverseπ/2 pulses magnetization,create transverse to evolve during t under the combined effects of the Zeeman interaction and indirect spin–spin couplings. magnetization, 2to evolve during t2 under the combined effects of the Zeeman interaction and indirect spin–spinDouble Fourier couplings. transformation Double Fourier of the transformation 2D dataset then of produces the 2D dataset a 2D plot then in whichproduces the regulara 2D plot MAS in NMR spectrum (detected during t ) is correlated with the multiplet (a doublet for a two-spin system), which the regular MAS NMR spectrum2 (detected during t2) is correlated with the multiplet (a doublet forarising a two-spin from indirect system), spin–spin arising from coupling indirect monitored spin–spin during coupling the incremented monitored during evolution the time incrementedt1. evolution time t1.

Molecules 2020, 25, 1400 11 of 39 Molecules 2019, 24, x FOR PEER REVIEW 11 of 40

Figure 8. (a) Schematic representation of the pulse sequence for heteronuclear-J-resolved spectroscopy. Figure 8. (a) Schematic representation of the pulse sequence for heteronuclear-J-resolved (b) A typical two-dimensional (2D) 11B{31P} J-resolved MAS NMR spectrum of a B/P FLP. In the direct spectroscopy. (b) A typical two-dimensional (2D) 11B{31P} J-resolved MAS NMR spectrum of a B/P dimension, the horizontal projection is equal to the 11B MAS NMR spectrum, while the spectrum FLP. In the direct dimension, the horizontal projection is equal to the 11B MAS NMR spectrum, while obtained from Fourier transforming the data in the indirect dimension shows the doublet splitting due the spectrum obtained from Fourier transforming the data in the indirect1 dimension shows the to J-coupling. The dipolar evolution time stated in the text is t1/2 = τ + 2 t180◦ where t180◦ is the 180◦ doublet splitting due to J-coupling. The dipolar evolution time stated in× the text is t1/2 = τ + ½ × t180° pulse length. where t180° is the 180° pulse length. J-Based Correlation Spectroscopy: Several different NMR experiments were designed to correlate J-Based Correlation Spectroscopy: Several different NMR experiments were designed to heteronuclear NMR resonances exploiting J-couplings, of which the insensitive nuclei enhanced by correlate heteronuclear NMR resonances exploiting J-couplings, of which the insensitive nuclei polarization transfer (INEPT) sequence proved to be well suited for FLP macrocycles offering sufficiently enhanced by polarization transfer (INEPT) sequence proved to be well suited for FLP macrocycles large isotropic J-couplings. The INEPT experiment is widely used as a signal amplification technique offering sufficiently large isotropic J-couplings. The INEPT experiment is widely used as a signal in liquid-state NMR, but it is equally effective under MAS conditions [56]. The sequence depicted in amplification technique in liquid-state NMR, but it is equally effective under MAS conditions [56]. Figure9a begins with a Hahn-echo at the non-observed nucleus. A simultaneous 180 ◦ pulse on the The sequence depicted in Figure 9a begins with a HAHN-echo at the non-observed nucleus. A observed channel recouples the isotropic indirect dipolar interaction, while the direct dipolar coupling simultaneous 180° pulse on the observed channel recouples the isotropic indirect dipolar interaction, is averaged out by MAS as τ1 typically exceeds several rotor periods. Thus, evolution during τ1 while the direct dipolar coupling is averaged out by MAS as 𝜏 typically exceeds several rotor proceeds exclusively due to J-coupling. The optimum timing for maximizing I–S polarization transfer periods. Thus, evolution during 𝜏 proceeds exclusively due to J-coupling. The optimum timing for via the two simultaneous 90◦ pulses on both channels is τ1 = τ2 = 1/4J. The transferred polarization maximizing I–S polarization transfer via the two simultaneous 90° pulses on both channels is τ1 = τ2 can either be detected on the insensitive (see Figure9) or on the more sensitive nucleus, which forms = 1/4J. The transferred polarization can either be detected on the insensitive (see Figure 9) or on the the polarization source (indirect detection), resulting in a considerable sensitivity enhancement. In the more sensitive nucleus, which forms the polarization source (indirect detection), resulting in a two-dimensional version, a heteronuclear correlation plot is observed in which covalently bonded considerable sensitivity enhancement. In the two-dimensional version, a heteronuclear correlation nuclei give rise to cross-peaks. Alternatively, J-based heteronuclear correlation spectra can also be plot is observed in which covalently bonded nuclei give rise to cross-peaks. Alternatively, J-based obtained via heteronuclear double quantum spectroscopy applied under standard MAS conditions [57] heteronuclear correlation spectra can also be obtained via heteronuclear double quantum (for more information on double-quantum (DQ) spectroscopy, see Section 2.2.4). In that pulse sequence, spectroscopy applied under standard MAS conditions [57] (for more information on double-quantum given by 90◦(S)–(1/2J)–90◦(I)–t1–90◦(I)–acquire(S, t2), S–I double quantum coherence is first excited by (DQ) spectroscopy, see Section 2.2.4). In that pulse sequence, given by 90°(S)–(1/21 J)–90°(I)–t1–90°(I)– the pair of 90◦(S) and 90◦(I) pulses, which must be separated by the time period 2 J. This DQ coherence acquire(S, t2), S–I double quantum coherence is first excited by the pair of 90°(S) and 90°(I) pulses, is allowed to evolve for an incremented evolution time t1, before it is stopped by a second 90◦ pulse which must be separated by the time period ½ J. This DQ coherence is allowed to evolve for an applied to one of the nuclear species (I). This pulse also triggers the detection period, where the regular incremented evolution time t1, before it is stopped by a second 90° pulse applied to one of the nuclear free induction decay of the S nuclei is being acquired. As the double quantum coherence evolves with species (I). This pulse also triggers the detection period, where the regular free induction decay of the the sum of the chemical shifts of the I and S nuclei involved, this simple sequence, equipped with the S nuclei is being acquired. As the double quantum coherence evolves with the sum of the chemical appropriate phase cycling schemes for coherence selection, can be used under regular MAS conditions shifts of the I and S nuclei involved, this simple sequence, equipped with the appropriate phase for indirect detection of insensitive nuclei, spectral editing via heteronuclear double quantum filtration, cycling schemes for coherence selection, can be used under regular MAS conditions for indirect and two-dimensional correlation spectroscopy [57]. The analogous homonuclear version is called detection of insensitive nuclei, spectral editing via heteronuclear double quantum filtration, and two- INADEQUATE (Incredible Natural Abundance Double Quantum Transfer Experiment)[58]. Although it dimensional correlation spectroscopy [57]. The analogous homonuclear version is called enjoys widespread use in other areas of materials science, so far, no such experiments were applied to INADEQUATE (Incredible Natural Abundance Double Quantum Transfer Experiment) [58]. Although it borane–phosphane systems, presumably as the detection of P—P bond connectivity is usually not an enjoys widespread use in other areas of materials science, so far, no such experiments were applied issue in FLP chemistry. to borane–phosphane systems, presumably as the detection of P---P bond connectivity is usually not an issue in FLP chemistry.

Molecules 2020, 25, 1400 12 of 39 Molecules 2019, 24, x FOR PEER REVIEW 12 of 40

Figure 9. One-dimensionalOne-dimensional (1D) (1D) ( (aa)) and and 2D 2D ( (bb)) refocused refocused insensitive insensitive nuclei nuclei enhanced enhanced by by polarization transfer (INEPT) experiment used used for polarization tr transferansfer developed via J-coupling to detect heteronuclear correlations. 2.2.4. Direct Spin–Spin (Dipolar) Coupling 2.2.4. Direct Spin–Spin (Dipolar) Coupling Basic Principles: Nuclear precession frequencies are further influenced by magnetic dipole–dipole Basic Principles: Nuclear precession frequencies are further influenced by magnetic dipole– interactions, as the magnetic moments of neighboring spins create local magnetic fields that affect dipole interactions, as the magnetic moments of neighboring spins create local magnetic fields that the precession frequencies of the probe nuclei. These depend both on distances and on orientations affect the precession frequencies of the probe nuclei. These depend both on distances and on (defined by the angle of the internuclear distance vector relative to the magnetic field direction). Again, orientations (defined by the angle of the internuclear distance vector relative to the magnetic field there is a homonuclear contribution describing interactions of the observed nuclei with spins of the direction). Again, there is a homonuclear contribution describing interactions of the observed nuclei same kind and a heteronuclear contribution describing interactions of the observed nuclei with spins with spins of the same kind and a heteronuclear contribution describing interactions of the observed of different kinds. Their corresponding Hamiltonians relevant for the NMR line shape in the limit of nuclei with spins of different kinds. Their corresponding Hamiltonians relevant for the NMR line first-order perturbation theory are given by shape in the limit of first-order perturbation theory are given by homo 2 1 ˆ = ˆ ˆ 1 ˆ ˆ + ˆ ˆ 𝐻 H=DD −𝑑,i,j (3𝑐𝑜𝑠di,j 3𝜃−1cos θ) 𝐼1 𝐼Iz,i −Iz,j 𝐼 I𝐼x,i I+𝐼x,j I𝐼y,i I,y,j ,(11) (11) ,, , − − , , 2− ,2 , , , 2 h i 𝜇𝛾µ𝛾γℏγ hetero 2 0 i j} 𝐻,,Hˆ = −𝑑=, 𝐼di,,j 𝐼Iˆ,z,iI(ˆz3𝑐𝑜𝑠,j 3cos𝜃−1θ ) 1 withwith 𝑑,d=i,j = , ,(12) (12) DD,i,j − − 4𝜋𝑟4πr3 where the operators 𝐼, , 𝐼, , and 𝐼, and 𝐼, , 𝐼, , and 𝐼, stand for the cartesian spin angular where the operators Iˆz,i, Iˆx,i, and Iˆy,i and Iˆz,j, Iˆx,j, and Iˆy,j stand for the cartesian spin angular momentum momentum components of the nuclei i and j. As indicated by Equation (12), the orientation components of the nuclei i and j. As indicated by Equation (12), the orientation dependence results in dependence results in line-broadening effects for polycrystalline samples. The dipolar coupling line-broadening effects for polycrystalline samples. The dipolar coupling constant di,j is proportional constant di,j is proportional to the inverse cube of the internuclear distance r, providing a to the inverse cube of the internuclear distance r, providing a straightforward connection to geometric straightforward connection to geometric structure. structure. Recoupling Heteronuclear Dipole Interactions: The most important experimental technique for Recoupling Heteronuclear Dipole Interactions: The most important experimental technique for measuring heteronuclear dipolar interactions under the high-resolution conditions of MAS NMR is measuring heteronuclear dipolar interactions under the high-resolution conditions of MAS NMR is the the rotational echo double resonance (REDOR) experiment [59–61]. As illustrated in Figure 10, MAS rotational echo double resonance (REDOR) experiment [59–61]. As illustrated in Figure 10, MAS causes causes the dipolar coupling to oscillate during the rotor period TR leading to its cancellation over a the dipolar coupling to oscillate during the rotor period TR leading to its cancellation over a complete complete rotor cycle. However, if the sign of the dipolar Hamiltonian is inverted by applying a π- rotor cycle. However, if the sign of the dipolar Hamiltonian is inverted by applying a π-pulse to the pulse to the non-observed I-spins, this average is non-zero and the interaction is said to be re-coupled. non-observed I-spins, this average is non-zero and the interaction is said to be re-coupled.

Molecules 2020, 25, 1400 13 of 39 Molecules 2019, 24, x FOR PEER REVIEW 13 of 40

Figure 10.10. Principle of the rotationalrotational echoecho doubledouble resonanceresonance (REDOR)(REDOR) experiment.experiment. (a) SchematicSchematic representation ofof the the dipolar dipolar interaction interaction of of two two spins spin I ands I and S inside S inside a spinning a spinning rotor rotor inclined inclined at the magicat the

angle.magic Localangle. magnetic Local magnetic fields produced fields pr byoduced their magneticby their magnetic moments (momentsBI and BS )(B modifyI and B theS) modify precession the frequencyprecession of frequency the interacting of the nuclei.interacting (b) Under nuclei. MAS (b) Under conditions, MAS the conditions, heteronuclear the heteronuclear dipolar Hamiltonian dipolar ˆ IS 𝐻 HHamiltonianD oscillates sinusoidally oscillates with sinusoidally rotor orientation, with leadingrotor orientation, to cancellation leading upon completionto cancellation of the upon rotor cyclecompletion in the fastof the spinning rotor cycle limit. in Sign the fast inversion spinning created limit. by Sign a π -pulseinversion applied created to the by non-observeda π-pulse applied I spins to (turningthe non-observedÎz into Îz I) spins interferes (turning with Î thisz into cancellation −Îz) interferes (bottom), with this resulting cancellation in the dipolar(bottom), re-coupling resulting einff ect.the − dipolar re-coupling effect. One then measures the normalized difference signal ∆S/So = (So S)/So in the absence (intensity − So) andOne the then presence measures (intensity the normalizedS) of the difference recoupling signal pulses. ΔS/S Theo = (S signalo − S)/S ofo in the the observe absence nuclei (intensity S is usuallySo) and the detected presence by (intensity a rotor-synchronized S) of the recoupling Hahn echo,pulses. while The signal the dipolar of the re-couplingobserve nuclei is eSff isected usually by 180detected◦ pulses by a during rotor-synchronized the rotor period, Hahn applied echo, while to the th Ie nuclei.dipolar Byre-coupling measuring is effected∆S/So as by a function180° pulses of dipolarduring the evolution rotor period, time NT appliedR, i.e., to the the duration I nuclei. of By one measuring rotor period ΔS/S multipliedo as a function by theof dipolar number evolution of rotor cycles,time NT theR, i.e., strength the duration of the dipole–dipole of one rotor period coupling multipli can beed characterized. by the number Pulse of rotor sequence cycles, imperfections the strength andof the finite dipole–dipole pulse length coupling effects can can be be characterized. corrected by thePulse compensation sequence imperfections scheme shown and in finite Figure pulse 11 (bottom)length effects [62]. can In this be method,corrected an by additional the compensationπ(I) pulse scheme applied shown simultaneously in Figure with11 (bottom) the π(S) [62]. refocusing In this pulsemethod, eliminates an additional the re-coupling π(I) pulse eappliedffect. This simultaneously “dummy” sequence with the can π(S) be refocusing considered pulse a more eliminates realistic referencethe re-coupling experiment effect. than This the standard“dummy” rotor sequence synchronized can be echo considered sequence, a evenmore though realistic the reference inclusion ofexperiment a third set than of measurements the standard rotor tends synchronized to reduce overall echo sequence, signal-to-noise even though ratio and thethere inclusion seems of toa third be a tendencyset of measurements to “over-correct” tends atto longerreduce evolutionoverall signal-t timeso-noise [62]. ratio and there seems to be a tendency to 1 “over-correct”In the case at oflongerI > 2evolutionnuclei, thetimes S{I} [62]. REDOR experiment is complicated by the quadrupolar interaction,In the whichcase of produces I > ½ nuclei, large o ffthe-resonance S{I} REDOR shifts experiment for the nuclei is involved complicated in the by non-central the quadrupolar Zeeman transitions,interaction, whichwhich in produces this case dolarge not off-resonance resonate with shifts the applied for the REDOR nucleiπ involvedrecoupling in pulses.the non-central For such cases,Zeeman the transitions,rotational echo which adiabatic in this passage case do double not resonate resonance with(REAPDOR) the applied [63 REDOR,64] technique π recoupling is very pulses. useful. ThisFor such method cases, replaces the rotational the π recoupling echo adiabatic pulses passage by continuous-wave double resonance irradiation (REAPDOR) of the [63,64] quadrupolar technique nuclei is duringvery useful. a duration This ofmethodTR/3 [replaces65]. As the the energy π recoupling levels of pulses the nuclei by continuous-wave in the non-central Zeemanirradiation states of the are modulatedquadrupolar by nuclei MAS, theyduring can a comeduration into resonanceof 𝑇/3 [65]. temporarily As the energy while thelevels radiofrequency of the nuclei field in the is on non- for thecentral quadrupolar Zeeman states nucleus, are leadingmodulated to population by MAS, they transfers can come between into these resonance levels temporarily during the irradiation while the period.radiofrequency In this field manner, is on dipolar for the quadrupolar interactions ofnucleus, the observe leading nuclei to population with the transfers quadrupolar between nuclei these in non-centrallevels during Zeeman the irradiation states are period. being In re-coupled this manner, as well,dipolar enhancing interactions the overallof the observe dipolar nuclei dephasing with etheffect quadrupolar in a calculable nuclei fashion. in non-ce Approximatentral Zeeman analytical states expressions are being re-coupled for short mixing as well, periods enhancing were alsothe givenoverall [66 dipolar]. dephasing effect in a calculable fashion. Approximate analytical expressions for short mixingRecoupling periods were Homonuclear also given Dipolar[66]. Interactions: For detecting homonuclear dipolar couplings under MAS conditions, a frequently utilized approach is the excitation of double-quantum (DQ) coherences. Such coherences involve the combined excitation of two simultaneous ∆m = 1 transitions for two nuclear spins. This effective ∆m = 2 transition can only occur if the two nuclear spins are coupled to each other. While both through-space and indirect spin–spin coupling can be exploited for DQ excitation, the former is predominantly utilized to obtain structural information in FLP chemistry.

Molecules 2020, 25, 1400 14 of 39

The simplest DQ excitation scheme involves the application of trains of two short 90◦ pulses (back-to-back (BaBa) sequence [67,68] (see Figure 12)); however, alternative, more sophisticated excitation schemes basedMolecules on 2019 gamma-encoding, 24, x FOR PEER REVIEW are also known [69–71]. 14 of 40

Figure 11. (a) Pulse sequence elements of the standard compensated REDOR experiment. The evolution Figure 11. (a) Pulse sequence elements of the standard compensated REDOR experiment. The of the dipolar Hamiltonian is shown below each pulse sequence element. Top: spin-echo experiment evolution of the dipolar Hamiltonian is shown below each pulse sequence element. Top: spin-echo on the observed S-channel to acquire the reference intensity S0. Middle: echo experiment with dipolar experiment on the observed S-channel to acquire the reference intensity S0. Middle: echo experiment recoupling, resulting in the intensity S. Bottom: “dummy” echo experiment producing a correction S’ with dipolar recoupling, resulting in the intensity S. Bottom: “dummy” echo experiment producing a tocorrection the dephased S’ to the intensity dephased to account intensity for to account pulse inaccuracies for pulse inaccuracies and ﬁnite and pulse finite length pulse e lengthﬀects. effects. (b) Pulse sequence(b) Pulse for sequence the rotational for the echo rotational adiabatic echo passage adiabatic double passage resonance double (REAPDOR) resonance experiment (REAPDOR) used forexperiment recoupling used of the for direct recoupling dipolar of interaction the direct todi quadrupolarpolar interaction nuclei. to quad Here,rupolar an adiabatic nuclei. transferHere, an (AT) pulseadiabatic inverts transfer magnetization (AT) pulse on invert the I-channel.s magnetization on the I-channel.

DQRecoupling excitation Homonuclear can be exploited Dipolar by Interactions: proving spatial For detecting proximity homonuclear between nuclei dipolar within couplings distinct, spectroscopicallyunder MAS conditions, resolved a sitesfrequently [69]. The utilized pulse approach sequence is in the this excitation case contains of double-quantum an evolution period (DQ) t1, duringcoherences. which Such the DQ coherences coherence involve evolves, the combined followed byexcitation reconversion of two tosimultaneous single-quantum Δm = (SQ)1 transitions coherence byfor a 90two◦ pulse nuclear after spins. which This the effective regular Δ MASm = 2 NMR transition signal can is monitoredonly occur duringif the two the nuclear detection spins period are t2. Separationcoupled to of each SQ and other. DQ While coherences both through-space is easily achieved and byindirect appropriate spin–spin phase coupling cycling. can As be DQ exploited coherence for DQ excitation, the former is predominantly utilized to obtain structural information in FLP chemistry. The simplest DQ excitation scheme involves the application of trains of two short 90°

Molecules 2019, 24, x FOR PEER REVIEW 15 of 40 pulses (back-to-back (BaBa) sequence [67,68] (see Figure 12)); however, alternative, more sophisticated excitation schemes based on gamma-encoding are also known [69–71].

Molecules 2020, 25, 1400 15 of 39

Moleculesevolves 2019 with, 24 the, x FOR sum PEER of the REVIEW involved resonance frequencies of both respective spins, the corresponding15 of 40 2D spectrum can serve to correlate the resonance frequencies of proximal spins, leading to oﬀ-diagonal pulsescross-peaks (back-to-backbetween (BaBa neighboring) sequence inequivalent [67,68] (see species Figure and12)); diagonalhowever,autocorrelation alternative, more peaks sophisticatedfor proximal excitationchemically schemes equivalent based spins. on gamma-encoding are also known [69–71].

Figure 12. Top: 2D double quantum/single quantum NMR spectroscopy with back-to-back (BaBa) excitation. All pulses have a 90° flip angle. Bottom: BaBa-XY16 excitation for increased efficiency and excitation bandwidth. The symbols x and y denote pulse phases.

DQ excitation can be exploited by proving spatial proximity between nuclei within distinct, spectroscopically resolved sites [69]. The pulse sequence in this case contains an evolution period t1, during which the DQ coherence evolves, followed by reconversion to single-quantum (SQ) coherence by a 90° pulse after which the regular MAS NMR signal is monitored during the detection period t2. Separation of SQ and DQ coherences is easily achieved by appropriate phase cycling. As DQ coherence evolves with the sum of the involved resonance frequencies of both respective spins, the correspondingFigureFigure 12. 12. Top:2DTop: spectrum2D 2D double double canquantum/single quantum serve /singleto correlate quantum quantum the NMR NMR resonance spectroscopy spectroscopy frequencies with with back-to-backback-to-back of proximal (BaBa(BaBa spins,) ) leadingexcitation.excitation. to off-diagonal All All pulses pulses have have cross-peaks a a 90° 90◦ flipflip angle. angle.between Bottom: Bottom: neighboring BaBa-XY16 BaBa-XY16 excita excitationinequivalenttion for for increased increased species efficiency e ffiandciency diagonaland and autocorrelationexcitationexcitation bandwidth.peaks bandwidth. for proximal The The symbols symbols chemically xx andand yy denoteequivalentdenote pulse pulse spins.phases. phases. The DQ signal intensity depends on both the strength of the magnetic dipole–dipole coupling The DQ signal intensity depends on both the strength of the magnetic dipole–dipole coupling and theDQ totalexcitation length can of thebe excitationexploited byperiod. proving Meas spatialurements proximity of such between DQ coherence nuclei build-upwithin distinct, curves and the total length of the excitation period. Measurements of such DQ coherence build-up curves spectroscopicallycan be used for the resolved quantitative sites [69].characterization The pulse sequence of spin systems in this case and contains distance an measurements evolution period [68]. tA1, can be used for the quantitative characterization of spin systems and distance measurements [68]. duringspecific which variant the of DQ this coherence principle evolves,is the DRENAR followed (dipolar by reconversion re-coupling to effects single-quantum nuclear alignment (SQ) coherence reduction) A specific variant of this principle is the DRENAR (dipolar re-coupling effects nuclear alignment reduction) bypulse a 90° sequence, pulse after which which monitors the regular an intensity MAS NMR difference signal isin monitored analogy to during the REDOR the detection method period [72–75]. t2. pulse sequence, which monitors an intensity difference in analogy to the REDOR method [72–75]. SeparationDRENAR measuresof SQ and the DQ decrease coherences of longitudinal is easily achieved magnetization by appropriate caused by phase the build-upcycling. Asof DQ DRENAR measures the decrease of longitudinal magnetization caused by the build-up of DQ intensity. coherenceintensity. Theevolves standard with thesequence sum of depicted the involved in Figu resonancere 13 consists frequencies of two of consecutive both respective DQ excitationspins, the The standard sequence depicted in Figure 13 consists of two consecutive DQ excitation blocks, followed correspondingblocks, followed 2D by spectruma 90° detection can serve pulse tofor correlate the residual the magnetization.resonance frequencies For the reference of proximal signal, spins, the by a 90◦ detection pulse for the residual magnetization. For the reference signal, the second block is leadingsecond blockto off-diagonal is phase-shifted cross-peaks by 90° (C’), between leading neighboring to inversion ofinequivalent the homonuclear species Hamiltonian and diagonal and, phase-shifted by 90◦ (C’), leading to inversion of the homonuclear Hamiltonian and, hence, cancellation autocorrelationhence, cancellation peaks offor the proximal DQ coherence. chemically In equivalentthe standard spins. DRENAR sequence [72,73], the difference of the DQ coherence. In the standard DRENAR sequence [72,73], the difference intensity (without and intensityThe DQ(without signal and intensity with dependsrecoupling) on bothis measured the strength as a offunction the magnetic of the dipole–dipoleexcitation time. coupling In the with recoupling) is measured as a function of the excitation time. In the constant-time (CT) variant, the andconstant-time the total length (CT) ofvariant, the excitation the phase period. shift Meas of theurements second of blocksuch DQrelative coherence to the build-up first block curves is phase shift of the second block relative to the first block is systematically incremented rather than the cansystematically be used for incremented the quantitative rather characterization than the DQ excitation of spin systems time [74]. and distance measurements [68]. A DQ excitation time [74]. specific variant of this principle is the DRENAR (dipolar re-coupling effects nuclear alignment reduction) pulse sequence, which monitors an intensity difference in analogy to the REDOR method [72–75]. DRENAR measures the decrease of longitudinal magnetization caused by the build-up of DQ intensity. The standard sequence depicted in Figure 13 consists of two consecutive DQ excitation blocks, followed by a 90° detection pulse for the residual magnetization. For the reference signal, the second block is phase-shifted by 90° (C’), leading to inversion of the homonuclear Hamiltonian and, hence, cancellation of the DQ coherence. In the standard DRENAR sequence [72,73], the difference intensity (without and with recoupling) is measured as a function of the excitation time. In the constant-time (CT) variant, the phase shift of the second block relative to the first block is systematically incremented rather than the DQ excitation time [74].

Figure 13. (a) Dipolar re-coupling eﬀects nuclear alignment reduction (DRENAR) (top) and constant-time (CT)-DRENAR (bottom) experiments based on reduction of longitudinal magnetization

by excitation of double quantum (DQ) coherence using the homonuclear dipolar coupling interaction. While, in the standard DRENAR experiment, the indirect dimension is incremented by adding DQ excitation blocks, in CT-DRENAR, the relative phase of the second excitation block is incremented by an arbitrary angle ϕ.(b) Exemplary CT-DRENAR spectrum for one phosphorus species in an FLP macrocycle further discussed in Section 3.6. Additionally, simulated CT-DRENAR curves for internuclear distances of 4.8, 5.3, and 5.8 Å are depicted.

Molecules 2020, 25, 1400 16 of 39

2.3. Cross-Polarization Finally, the magnetic dipole–dipole coupling is also exploited in cross-polarization (CP), in which polarization of a high-γ abundant spin system (usually protons) is transferred to recipient heteronuclei whose enhanced signal is then detected. To achieve an energy-conserving heteronuclear magnetization transfer, an identical Zeeman splitting for both isotopes involved is required, corresponding to equal precession frequencies of both nuclear species I and S. This is only possible in the doubly rotating frame, with radiofrequency pulses applied to both nuclei [76] under spin-locking conditions, adjusting both radiofrequency amplitudes according to the MAS-modulated Hartmann–Hahn matching condition [77] (see Figure 14). ν n νR = ν , (13) 1,I ± × 1,S where ν1,I and ν1,S are the nutation frequencies of the I and S nuclei, νR is the rotor frequency, and n is an integer. A part of the enhanced detection sensitivity arises from the much shorter spin-lattice relaxation times of the 1H source nuclei compared to the recipient heteronuclei. In fact, this is the main reason why 31P MAS NMR spectra on FLPs are usually obtained via cross-polarization from 1H spin reservoirs. The key to the cross-polarization principle is the fact that the abundant-spin system in the spin-locked state is far away from equilibrium conditions and tends toward equilibrium by losing magnetization. While this naturally occurs due to spin-lattice relaxation (time constant T1ρ, Figure 14c), the double irradiation experiment under the condition in Equation (12) opens up an alternative cross-relaxation path, via which the magnetization is channeled to the heteronuclei utilizing the magnetic dipole–dipole coupling mechanism. This transfer is characterized by the transfer time constant TCP, which depends on the strength of the heteronuclear dipolar coupling. Owing to the interplay of these competing relaxation channels, the achievable signal enhancement will depend on the duration of the double irradiation period (the so-called contact time, tct) which must be optimized experimentallyMolecules 2019, 24 on, x FOR the PEER sample. REVIEW 17 of 40

FigureFigure 14. 14.Principle Principle of theof the cross-polarization cross-polarization (CP) (CP) NMR NMR experiment: experiment: ( a()a) Zeeman Zeeman states states for for an an I (orange)I and(orange) an S nucleus and an (blue)S nucleus along (blue) the along quantization the quantizatio axis,n splitaxis, bysplit the by externalthe external magnetic magnetic field field (left) (left) and by and by the magnetic component B1 of the applied radiofrequency field under spin-lock conditions. the magnetic component B1 of the applied radiofrequency field under spin-lock conditions. Here, full circlesHere, illustrate full circles population illustrate dipopulationfferences. differences. (b) Pulse sequence (b) Pulse timingsequence diagram. timing diagram. During theDuring spin-lock the time, spin-lock time, the nutation frequency of one of the nuclear species is ramped to account for the nutation frequency of one of the nuclear species is ramped to account for modulation effects in the modulation effects in the Hartmann–Hahn matching condition caused by MAS. (c) Competitive cross- Hartmann–Hahn matching condition caused by MAS. (c) Competitive cross-relaxation and rotating relaxation and rotating frame spin-lattice relaxation processes. (d) Signal intensity as a function of frame spin-lattice relaxation processes. (d) Signal intensity as a function of contact time resulting from contact time resulting from these competitive processes. these competitive processes. 3. Results and Discussion

3.1. NMR Criteria of FLP Behavior Specific characteristics of vicinal borane–phosphane frustrated Lewis pairs (B/P FLPs) as determined experimentally and confirmed by DFT calculations were previously reviewed [19,78–80]. Both FLP centers are mostly characterized by an “in-between” coordination state between three and four, owing to the partial covalent bond between the Lewis centers, affecting the 11B chemical shifts and the 11B nuclear electric quadrupolar coupling constants. Figure 15 illustrates that both parameters are well correlated with each other, indicating that the strength of this partial bond dominates the trends in both NMR observables [79]. In addition, these parameters are well-correlated with the B…P internuclear distance, which varies between 2.00 Å (strongly interacting) and 2.20 Å (weakly interacting Lewis pairs). As a feature common to all vicinal B/P FLPs studied so far, the principal axis of the 11B EFG tensor makes an angle of ~21 ± 3° with the B---P vector [19]. The covalent interaction also manifests itself in 11B–31P indirect spin–spin couplings with an isotropic coupling constant of up to 60 Hz. 11B{31P} REDOR experiments further indicate a contribution from a J-anisotropy ΔJ on the order of ~100 Hz.

Molecules 2020, 25, 1400 17 of 39

3. Results and Discussion

3.1. NMR Criteria of FLP Behavior Specific characteristics of vicinal borane–phosphane frustrated Lewis pairs (B/P FLPs) as determined experimentally and confirmed by DFT calculations were previously reviewed [19,78–80]. Both FLP centers are mostly characterized by an “in-between” coordination state between three and four, owing to the partial covalent bond between the Lewis centers, affecting the 11B chemical shifts and the 11B nuclear electric quadrupolar coupling constants. Figure 15 illustrates that both parameters are well correlated with each other, indicating that the strength of this partial bond dominates the trends in both NMR observables [79]. In addition, these parameters are well-correlated with the B ... P internuclear distance, which varies between 2.00 Å (strongly interacting) and 2.20 Å (weakly interacting Lewis pairs). As a feature common to all vicinal B/P FLPs studied so far, the principal axis of the 11B EFG tensor makes an angle of ~21 3 with the B—P vector [19]. The covalent interaction ± ◦ also manifests itself in 11B–31P indirect spin–spin couplings with an isotropic coupling constant of up to 60 Hz. 11B{31P} REDOR experiments further indicate a contribution from a J-anisotropy ∆J on the Moleculesorder of 2019 ~100, 24, Hz.x FOR PEER REVIEW 18 of 40

11 11 FigureFigure 15. 15. CorrelationCorrelation of of 11BB isotropic isotropic chemical chemical shifts shifts and and 11BB nuclear nuclear electric electric quadrupole quadrupole coupling coupling 2 constants for vicinal B/P FLPs [78,79]. Closed symbols, solid line (C = 1.652 MHz + 0.040 δ/ppm; R2 constants for vicinal B/P FLPs [78,79]. Closed symbols, solid line (CQQ = 1.652 MHz + 0.040 ×× δ/ppm; R = 0.84): experimental data; open symbols, dashed line (C = 1.738 MHz + 0.040 δ/ppm; R2 = 0.87): = 0.84): experimental data; open symbols, dashed line (CQQ = 1.738 MHz + 0.040 ×× δ/ppm; R2 = 0.87): DFT-calculatedDFT-calculated values. values. 3.2. Dimeric Structures 3.2. Dimeric Structures Molecular aggregation, oligomerization, and polymerization of FLPs are often observed as Molecular aggregation, oligomerization, and polymerization of FLPs are often observed as a a consequence of strong covalent intermolecular interactions between the monomers. In the case consequence of strong covalent intermolecular interactions between the monomers. In the case of the of the cyclic FLP 1 shown in Figure 16, crystallization produces a simple dimer [25]. A dimeric cyclic FLP 1 shown in Figure 16, crystallization produces a simple dimer [25]. A dimeric structure structure was also obtained for the adduct 2, which is prepared by reacting the polymeric FLP was also obtained for the adduct 2, which is prepared by reacting the polymeric FLP [Ph2P–CH=CH– [Ph2P–CH=CH–B(C6F5)2]n with benzaldehyde (the crystal structure 2 will be the subject of a B(C6F5)2]n with benzaldehyde (the crystal structure 2 will be the subject of a subsequent publication). subsequent publication). While the completely amorphous nature of the polymeric material precludes While the completely amorphous nature of the polymeric material precludes a conventional structural characterization using single crystals, solid-state NMR was able to develop important structural constraints [19,78]. An essential feature of B/P FLP aggregates concerns the structural analysis by REDOR and REAPDOR experiments, which can no longer be analyzed in terms of simple 11B–31P two-spin interactions. Rather, the multi-spin character of these interactions must be taken into account. This is illustrated in Figure 16b as an example of the 31P{11B} REAPDOR curves measured for 1 and 2.

Molecules 2020, 25, 1400 18 of 39 a conventional structural characterization using single crystals, solid-state NMR was able to develop important structural constraints [19,78]. An essential feature of B/P FLP aggregates concerns the structural analysis by REDOR and REAPDOR experiments, which can no longer be analyzed in terms of simple 11B–31P two-spin interactions. Rather, the multi-spin character of these interactions must be taken into account. This is illustrated in Figure 16b as an example of the 31P{11B} REAPDOR curves Moleculesmeasured 2019, 24 for, x 1FORand PEER2. REVIEW 19 of 40

Figure 16. (a) Schematic representation of B/P FLP dimer and dimeric adduct of a B/P FLP with Figure 16. (a) Schematic representation of B/P FLP dimer and dimeric adduct of a B/P FLP with benzaldehyde and (b) characterization of both compounds using 31P{11B} REAPDOR experiments. The benzaldehyde and (b) characterization of both compounds using 31P{11B} REAPDOR experiments. The complete simulation and its three isotopologue constituents are shown as summarized in the insets [79].complete simulation and its three isotopologue constituents are shown as summarized in the insets [79].

31 11 AsAs each each of of thethe 31PP nucleinuclei interacts interacts with with two two closest closestB 11 spins,B spins, the existencethe existence of di ffoferent different isotopologues 11 10 isotopologuesmust be considered, must be considered, reflecting thereflecting natural the abundances natural abundances of B andof 11B Band nuclei. 10B nuclei. For compound For 1, compoundthese contributions 1, these contributions consist of consist a dominant of a dominant three-spin three-spin component component (64.2%, (64.2%, distances distances 2.08 2.08 and 2.75 Å, 11 ... 31 ... 11 andB 2.75 Å,P 11B…31BP angle…11B angle of 90 of◦, 90°, blue blue curve) curve) and and two two minor minor two-spin two-spin components components fromfrom the two possible possible11B–31P– 11B–10B31P– isotopologues10B isotopologues (15.9% (15.9% each), each), while while 4% 4% of of the the moleculesmolecules do do not not contain contain any any 11B 11and,B and, thus, thus,will will yield yield no REAPDORno REAPDOR eff ect.effect. For For compound compound2 ,2 the, the situation situation is is completely completely analogous, except except that the that... the.. B…..P distances are much longer, i.e., 3.94 and 4.27 Å, respectively, with11 a 11B...…3131P…11...B 11angle of B P distances are much longer, i.e., 3.94 and 4.27 Å, respectively, with a B P B angle of 83.8◦. 83.8°.Figure Figure 16b illustrates16b illustrates these these individual individual contributions contributions and and the the corresponding corresponding predicted overall overall 31P{11B} 31 11 REAPDORP{ B} REAPDOR curves, curves, showing showing good good agreement agreement with with the the experimental experimental data data in inboth both systems.systems. In In the case the case of 1, the agreement can be improved even further if a J-anisotropy of 100 Hz is included in of 1, the agreement can be improved even further if a J-anisotropy of 100 Hz is included in the analysis. the analysis. Analogous considerations apply for the analysis of 11B{31P} REDOR data. Three-spin Analogous considerations apply for the analysis of 11B{31P} REDOR data. Three-spin 11B ... 31P ... 11B 11B…31P…11B REDOR and REAPDOR curves show a considerable dependence on the angle subtended byREDOR the two and internuclear REAPDOR vectors curves [79]. show In systems a considerable where the dependencedetails of stru onctural the geometry angle subtended are not by the known,two internuclear constraints for vectors such angles [79]. Incan, systems thus, bewhere developed the on details the basis of structural of experimental geometry NMR are data. not known, constraints for such angles can, thus, be developed on the basis of experimental NMR data. 3.3. FLP-Cyclotrimer 3.3. FLP-Cyclotrimer Treatment of (C6F5)2P–CH(CH3) –CH2–B(C6F5)2 with 9-borabicyclo[3.3.1]nonane (9-BBN) leads to the exchangeTreatment of H of for (C one6F5 C)26P–CH(CHF5 substituent3) –CH at the2–B(C borane6F5 )moiety2 with and 9-borabicyclo[3.3.1]nonane subsequent aggregation of (9-BBN) this leads intermediateto the exchange to the of cyclotrimer H for one 3 C depicted6F5 substituent in Figure at 17. the The borane solid-st moietyate NMR and parameters subsequent reveal aggregation strong of this intermolecular P-B covalent interactions, signifying relatively modest FLP reactivity [81]. The 11B{31P} REDOR and 31P{11B} REAPDOR data are consistent with the crystallographic distance of 2.006 Å, taking into account a J-anisotropy ΔJ of 100 Hz. The correct value of ΔJ is found by comparing the experimental data with a theoretical curve calculated on the basis of the crystallographically known

Molecules 2020, 25, 1400 19 of 39 intermediate to the cyclotrimer 3 depicted in Figure 17. The solid-state NMR parameters reveal strong intermolecular P-B covalent interactions, signifying relatively modest FLP reactivity [81]. The 11B{31P} REDOR and 31P{11B} REAPDOR data are consistent with the crystallographic distance of 2.006 Å, takingMolecules into 2019, account 24, x FOR aPEERJ-anisotropy REVIEW ∆J of 100 Hz. The correct value of ∆J is found by comparing20 of 40 the experimental data with a theoretical curve calculated on the basis of the crystallographically known internuclearinternuclear distances and and using using Δ∆JJ as as an an adjustable adjustable parameter. parameter. Alternatively, Alternatively, Δ∆J Jcancan be be calculated calculated fromfrom firstfirst principles. principles. In In general, general, both both approaches approaches are are found found in goodin good agreement agreement [19, 20[19,20].]. The The main main NMR observableNMR observable proving proving the cyclotrimeric the cyclotri structuremeric structure is the homonuclear is the homonuclear31P–31P dipole–dipole 31P–31P dipole–dipole interaction strengthinteraction defined strength by the defined trigonal by geometrythe trigonal with geometry P ... P distances with P… ofP distances 5.758(0) Å. of Figure 5.758(0) 17 bÅ. shows Figure that 17b the 31showsP–31P that DQ-DRENAR the 31P–31P curve DQ-DRENAR is quantitatively curve is consistent quantitatively with this consistent geometry. with For this a correct geometry. reproduction For a ofcorrect the experimental reproduction DRENAR of the experimental curve, the 31 DRENARP chemical curve, shift anisotropy the 31P chemical (CSA) mustshift anisotropy be accounted (CSA) for in themust simulation, be accounted as it for aff ectsin the the simulation, efficiency ofas dipolar it affects recoupling the efficiency [73]. of In dipo contrast,lar recoupling the intermolecular [73]. In … Pcontrast,... P distances the intermolecular between diff erentP P moleculardistances between units in monomericdifferent molecular P/B systems units are in significantly monomeric longer, P/B andsystems no DRENAR are significantly effect would longer, be and expected. no DRENAR effect would be expected.

Figure 17. (a) Formation of FLP-cyclotrimer 3 and (b) 31P–31P DRENAR build-up curve for 3. SolidFigure squares: 17. (a) Formation experimental of FLP-cyclotrimer data points. Solid3 and curves: (b) 31P–31 simulatedP DRENAR DRENAR build-up curve curve basedfor 3. Solid on the intermolecularsquares: experimental dipole–dipole data couplingpoints. onlySolid (lower curves: curve) simulated and with DRENAR additional curve consideration based ofon the the31 P CSAintermolecular (∆σ = 99 ppm), dipole–dipole leading to coupling an improved only agreement(lower curve) with and the with experimental additional data. consideration Reproduced of the from 31 Δ ReferenceP CSA ( [σ81 =]. 99 ppm), leading to an improved agreement with the experimental data. Reproduced from Reference [81].

3.4. Dimeric and Trimeric CO Adducts While the reaction of vicinal B/P FLPs with carbon monoxide results in cooperative binding producing simple cyclic adducts [82], the reaction products of trifunctional P/B/B FLPs offering two such borane Lewis acid centers and one phosphane Lewis base center allow for a rearrangement,

Molecules 2020, 25, 1400 20 of 39

3.4. Dimeric and Trimeric CO Adducts While the reaction of vicinal B/P FLPs with carbon monoxide results in cooperative binding producingMolecules 2019 simple, 24, x FOR cyclic PEER adducts REVIEW [82 ], the reaction products of trifunctional P/B/B FLPs oﬀering two21 suchof 40 borane Lewis acid centers and one phosphane Lewis base center allow for a rearrangement, leading leadingto the integration to the integration of the CO of moiety the CO into moiety the FLP into backbone. the FLP Suchbackbone. compounds Such compounds show a pronounced show a pronouncedtendency to self-assembletendency to self-assemble into dimeric and into trimeric dimeric macrocycles and trimeric (see macrocycles Figure 18), (see which Figure depends 18), which on the dependsnature of on the the substituents nature of the at the substi phosphanetuents at unit the phosphane [83]. unit [83].

Figure 18. Preparation routes and molecular structures of B/P FLP–CO dimeric and trimeric macrocycles. Figure 18. Preparation routes and molecular structures of B/P FLP–CO dimeric and trimeric (a) Synthesis of a B/P FLP with two borane Lewis centers; (b) reaction of this B/P FLP with CO, resulting macrocycles. (a) Synthesis of a B/P FLP with two borane Lewis centers; (b) reaction of this B/P FLP in a five-membered intermediate, which (c) can self-assemble to dimeric or trimeric adducts. with CO, resulting in a five-membered intermediate, which (c) can self-assemble to dimeric or trimeric adducts.The structural features of these supramolecular adducts were recently characterized by solid-state NMR [83]. Figure 19a shows the 11B MAS NMR spectra of both compounds. The resonance assignments The structural features of these supramolecular adducts were recently characterized by solid- to the two distinct boron species are easily accomplished with the help of DFT calculations for both the state NMR [83]. Figure 19a shows the 11B MAS NMR spectra of both compounds. The resonance 11B quadrupolar coupling and the magnetic shielding tensors. Compared to the FLP-like species (B1 and assignments to the two distinct boron species are easily accomplished with the help of DFT B3), the O-bonded boron nuclei (B2 and B4) are distinguished by significantly stronger EFGs and more calculations for both the 11B quadrupolar coupling and the magnetic shielding tensors. Compared to positive chemical shift values. In the 11B MAS NMR spectra, additional multiplicities are observed due the FLP-like species (B1 and B3), the O-bonded boron nuclei (B2 and B4) are distinguished by to slight crystallographic non-equivalences of the boron sites. For example, the 11B MAS NMR spectrum significantly stronger EFGs and more positive chemical shift values. In the 11B MAS NMR spectra, of the trimeric compound shows two groups of resonances, both containing contributions from three additional multiplicities are observed due to slight crystallographic non-equivalences of the boron different monomeric units (B2, B4, B6 for the O-bonded boron species and B1, B3, B5 for the FLP-like sites. For example, the 11B MAS NMR spectrum of the trimeric compound shows two groups of species) in equal proportions. In the 31P MAS NMR spectra (Figure 19b), the two crystallographically resonances, both containing contributions from three different monomeric units (B2, B4, B6 for the O- different phosphorus positions of the dimeric compound are barely resolved. Again, the spectra show bonded boron species and B1, B3, B5 for the FLP-like species) in equal proportions. In the 31P MAS evidence of 31P–11B indirect spin–spin coupling, revealing a multiplet structure. Thus, the simulation NMR spectra (Figure 19b), the two crystallographically different phosphorus positions of the dimeric must include separate contributions from the 10B and 11B isotopologues. This spectrum looks even compound are barely resolved. Again, the spectra show evidence of 31P–11B indirect spin–spin more complex for the trimeric compound, consisting of the three contributions (P1–P3) for each of the coupling, revealing a multiplet structure. Thus, the simulation must include separate contributions two possible isotopologues of the macrocycle and of a minor impurity. To simplify the complex overall from the 10B and 11B isotopologues. This spectrum looks even more complex for the trimeric lineshapes, the 1H 31P{11B} CP/MAS NMR spectra (Figure 19c) were also acquired for both samples compound, consisting→ of the three contributions (P1–P3) for each of the two possible isotopologues using swept-frequency two-pulse phase modulation SWFTPPM decoupling of 1H and 11B nuclei of the macrocycle and of a minor impurity. To simplify the complex overall lineshapes, the 1H31P{11B} CP/MAS NMR spectra (Figure 19c) were also acquired for both samples using swept- frequency two-pulse phase modulation SWFTPPM decoupling of 1H and 11B nuclei simultaneously [84]. This proved to be useful for developing additional constraints for the line-shape fitting procedure. The rather large indirect spin–spin coupling constants 1Jiso(11B–31P) measured for these macrocycles (70 Hz and 80 Hz for the dimer and the trimer, respectively) demonstrate that the dative B---P bonds are comparatively strong compared to other systems. This observation is consistent with

Molecules 2020, 25, 1400 21 of 39

simultaneously [84]. This proved to be useful for developing additional constraints for the line-shape 1 11 31 ﬁtting procedure. The rather large indirect spin–spin coupling constants Jiso( B– P) measured for Molecules 2019, 24, x FOR PEER REVIEW 22 of 40 theseMolecules macrocycles 2019, 24, x (70 FOR Hz PEER and REVIEW 80 Hz for the dimer and the trimer, respectively) demonstrate22 that of the40 dative B—P bonds are comparatively strong compared to other systems. This observation is consistent the comparatively small 11B nuclear electric quadrupolar coupling constants and rather negative 11B withthe the comparatively comparatively small small 11B11 nuclearB nuclear electric electric quadrupolar quadrupolar coupli couplingng constants constants and andrather rather negative negative 11B chemical shift values observed here in relation to other B/P FLPs. The 2D 11B{31P} J-resolved MAS 11B chemical shift values observed here in relation to other BB/P/P FLPs.FLPs. The 2D 11B{B{3131P}P} J-resolved J-resolved MAS MAS NMR experiments depicted in Figure 20 confirm the 11B resonance assignments and coupling NMRNMR experiments experiments depicted depicted in Figure in Figure 20 conﬁrm 20 confirm the 11B the resonance 11B resonance assignments assignments and coupling and coupling constants. constants. constants.

Figure 19. Experimental and simulated (a) 11B MAS, (b) 31P MAS, and (c) 1H31P{11B dec} CP/MAS FigureFigure 19. Experimental19. Experimental and and simulated simulated (a )(11a) B11 MAS,B MAS, (b ()b31) P31P MAS, MAS, and and ( c()c)1 H1H3131P{11BB dec} dec} CP/MAS CP/MAS NMR spectra of the B/P FLP–CO dimer 4 and trimer 5. Simulated 31P MAS NMR spectra→ include the NMRNMR spectra spectra of theof the B/P B/P FLP–CO FLP–CO dimer dimer4 and 4 and trimer trimer5. 5 Simulated. Simulated31 31PP MASMAS NMR spectra include include the the individual contributions for 11B (green) and 10B (red) isotopologues. Note the effect of 11B decoupling individualindividual contributions contributions for for11B 11 (green)B (green) and and10 B10B (red) (red) isotopologues. isotopologues. NoteNote the eeffectﬀect of 11BB decoupling decoupling suppressing the multiplet structure due to 31P–11B indirect spin–spin coupling in (c). Impurities are suppressingsuppressing the the multiplet multiplet structure structure due due to to31 P–31P–1111BB indirect indirect spin–spinspin–spin couplingcoupling in (c). Impurities Impurities are are marked with a cross. In (a) the signals labeled B7 and B8 originate from an impurity. markedmarked with with a cross. a cross. In (a)In (a) the the signals signals labeled labeled B7 B7 and and B8 B8 originate originate fromfrom anan impurity.impurity.

Figure 20. The 2D 11B{1131P} 31J-resolved MAS NMR spectra of (a) 4 and (b) 5. In (b), the additional sites FigureFigure 20. The20. The 2D 2DB{ 11B{P}31P} J-resolved J-resolved MAS MAS NMR NMR spectra spectra ofof ((aa)) 44 andand ((bb)) 5.. In ( (b),), the the additional additional sites sites B7 and B8 originate from an impurity. B7 andB7 and B8 originateB8 originate from from an an impurity. impurity.

Figure 21 shows the 2D heteronuclear correlation spectra for both compounds obtained using Figure 21 shows the 2D heteronuclear correlation spectra for both compounds obtained using the 11B{31P} CP/refocused INEPT MAS NMR experiment. For the dimeric macrocycle, both cross peaks the 11B{31P} CP/refocused INEPT MAS NMR experiment. For the dimeric macrocycle, both cross peaks

Molecules 2019, 24, x FOR PEER REVIEW 23 of 40 Molecules 2020, 25, 1400 22 of 39 between P1 and B1 and between P2 and B3 can be identified in Figure 21, while the spectrum of the Molecules 2019, 24, x FOR PEER REVIEW 23 of 40 trimeric compounds shows the expected three cross peaks (P1–B1, P2–B3, and P3–B5) with a good resolution.Figure Additionally, 21 shows the 2Da fourth heteronuclear cross peak correlation between spectra P4 and for B7 both corresponding compounds to obtained the impurity using the is 11between31 P1 and B1 and between P2 and B3 can be identified in Figure 21, while the spectrum of the B{ P} CP/refocused INEPT MAS NMR experiment. For the dimeric31 macrocycle, both cross peaks clearlytrimeric evident. compounds Note showsin particular the expected the improved three cross resolution peaks (P1–B1,in the P2–B3,P dimension and P3–B5) compared with ato good the between P131 and B1 and between P2 and B3 can be identified in Figure 21, while the spectrum of the standardresolution. P Additionally, MAS NMR spectra a fourth of Figurecross peak 19b. Furthermore,between P4 and the B72D correspondingaspect of this INEPT to the experimentimpurity is trimeric compounds shows the expected11 three cross peaks (P1–B1, P2–B3, and P3–B5) with a good11 allowsclearly improved evident. Noteresolution in particular in the Bthe dimension improved as resolution well. Since in the the cross 31P dimensionpeaks are well compared resolved, to theB MASresolution. spectra31 Additionally, of the single a fourth species cross can peak be obtained between P4by andanalyzing B7 corresponding the Fourier to transforms the impurity and is clearly their standard P MAS NMR spectra of Figure 19b. Furthermore,31 the 2D aspect of this INEPT experiment 11 31 separatedevident. Note rows in along particular the thedirect improved dimension.11 resolution Figure in 22 the showsP dimension the B{ comparedP} REDOR to results the standard of 115, 31allows improved resolution in the B dimension as well. Since the cross peaks are well resolved, B illustratingP MAS NMR a dramatic spectra difference of Figure 19betweenb. Furthermore, the B/P FL theP-like 2D aspectborane of unit this B1, INEPT B3, B5, experiment and B7 (closest allows MAS spectra of the single 11species can be obtained by analyzing the Fourier transforms and11 their improved… resolution in the B dimension as well. Since the cross… peaks are well resolved, B MAS PseparatedB distance rows 2.10 alongÅ) and the the directO-bonded dimension. boron species Figure (closest 22 shows P B thedistance 11B{31 4.38P} REDOR Å). Figure results 23 shows of 5 , spectra31 11 of the single species can be obtained by analyzing the Fourier transforms and their separated theillustrating P{ B} REAPDOR a dramatic data difference along with between their simulations.the B/P FLP-like As in borane the previous unit B1, examples, B3, B5, theand REAPDOR B7 (closest rows along the direct dimension. Figure 22 shows the 11B{31P} REDOR results of 5, illustrating a responseP…B distance must 2.10 be Å)considered and the O-bonded as the weighted boron species sum of (closest the four P… Bpossible distance isotopologues. 4.38 Å). Figure While 23 shows an dramatic difference between the B/P FLP-like borane unit B1, B3, B5, and B7 (closest P ... B distance approximatethe 31P{11B} REAPDOR analysis based data alongon the with closest their distance simulations. of 2.10 As Å alreadyin the pr comesevious reasonably examples, theclose, REAPDOR we note 2.10 Å) and the O-bonded boron species (closest P ... B distance 4.38 Å). Figure 23 shows the 31P{11B} thatresponse the agreement must be withconsidered the experimental as the weighted data is significantlysum of the fourimproved possible when isotopologues. including the While second- an REAPDOR data along with their simulations. As in the previous examples, the REAPDOR response nearestapproximate neighbor analysis (4.38 basedÅ) in theon theanalysis. closest distance of 2.10 Å already comes reasonably close, we note must be considered as the weighted sum of the four possible isotopologues. While an approximate that the agreement with the experimental data is significantly improved when including the second- analysis based on the closest distance of 2.10 Åalready comes reasonably close, we note that the nearest neighbor (4.38 Å) in the analysis. agreement with the experimental data is significantly improved when including the second-nearest neighbor (4.38 Å) in the analysis.

Figure 21. 11B{31P} CP/refocused INEPT MAS NMR spectra of (a) 4 and (b) 5. The sites P4 and B7 in (b) arise from an impurity.

FigureFigure 21.21. 1111B{31P}P} CP/refocused CP/refocused INEPT INEPT MAS MAS NMR NMR spectra spectra of of ( (aa)) 4 and ( b) 5. The sites P4 andand B7B7 inin

((bb)) arisearise fromfrom anan impurity.impurity.

FigureFigure 22. 22. ((aa)) Experimental Experimental 1111B{B{3131P}P} REDOR REDOR curves curves (squares) (squares) obtained obtained for for 55. .Data Data represented represented by by the the blackblack and and red red squares squares were were obtained by by line-shape deconvolution deconvolution of of every every 1D 1D REDOR REDOR (S) (S) and and echo echo spectrumspectrum (S (S00)) (see (see Figure Figure 11),11), using using a a single single line line for for each each of of the the two two boron boron environments, environments, i.e., i.e., for for boron boron Figure 22. (a) Experimental 11B{31P} REDOR curves (squares) obtained for 5. Data represented by the black and red squares were obtained by line-shape deconvolution of every 1D REDOR (S) and echo

spectrum (S0) (see Figure 11), using a single line for each of the two boron environments, i.e., for boron

Molecules 2020, 25, 1400 23 of 39

Molecules 2019, 24, x FOR PEER REVIEW 24 of 40

species bound to phosphorus (black) and oxygen (red). Additionally, the intensity of an integral over species bound to phosphorus (black) and oxygen (red). Additionally, the intensity of an integral over the total superimposed spectrum is depicted (blue squares). Corresponding simulations based on the the total superimposed spectrum is depicted (blue squares). Corresponding simulations based on the single-crystal structure are included as solid lines. The simulations consider a two-spin P–B system single-crystal structure are included as solid lines. The simulations consider a two-spin P–B system with a distance of 2.10 Å (black curve) and a three-spin 11B…31P…11B system with distances of 4.38 and with a distance of 2.10 Å (black curve) and a three-spin 11B ... 31P ... 11B system with distances of … … 5.4 Å and a B P B angle... of 117°... (red curve). Furthermore, the normalized sum of both simulations is 4.38 and 5.4 Å and a B P B angle of 117◦ (red curve). Furthermore, the normalized sum of both shown (blue curve). In (b), the REDOR (bottom) and the reference spectra (top) acquired for a dipolar simulations is shown (blue curve). In (b), the REDOR (bottom) and the reference spectra (top) acquired evolution time of 0.48 ms are depicted along with corresponding line-shape simulations. for a dipolar evolution time of 0.48 ms are depicted along with corresponding line-shape simulations.

FigureFigure 23. ExperimentalExperimental and and simulated 31P{P{1111B}B} REAPDOR REAPDOR curves curves obtained obtained for for 5.. Corresponding Corresponding simulationssimulations based based on on the the single-crystal single-crystal structure structure are are included included as as solid solid lines. lines. The The simulated simulated curve curve is is thethe weighted weighted sum sum of of the the individual individual curves curves obtained obtained for for the the three three poss possibleible isotopologues: isotopologues: a a three-spin three-spin 11 ... 31 ... 11 ...... 11BB…31P…11PB systemB system with distances with distances of 2.10 of and 2.10 4.38 and Å 4.38 (64.2%, Å (64.2%, red curve) red curve) and a andB…P… aB B angleP ofB 112° angle and of 11 ... 31 two112 ◦11andB…31 twoP two-spinB systemsP two-spin with systems distances with of 2.10 distances and 4.38 of 2.10Å (blue and and 4.38 green Å (blue curves). and green The dashed curves). curveThe dashed is an approximate curve is anapproximate two-spin simulation two-spin, neglecting simulation, the neglecting secondary the 4.38 secondary Å distance. 4.38 Å distance.

3.5. Tetrameric B/P FLP–CO2 Macrocycles. 3.5. Tetrameric B/P FLP–CO2 Macrocycles. The six-membered cyclic B/P FLP 6 shown in Figure 24 adds carbon dioxide under mild conditions. The six-membered cyclic B/P FLP 6 shown in Figure 24 adds carbon dioxide under mild Upon crystallization, a unique macrocyclic tetramer 7 with bridging CO2 molecules between the conditions. Upon crystallization, a unique macrocyclic tetramer 7 with bridging CO2 molecules individual cyclo-FLP units was obtained [85]. The spectroscopic features of this macrocycle were between the individual cyclo-FLP units was obtained [85]. The spectroscopic features of this compared with those of a closely related monomeric B/B/P FLP adduct 8, which can be stabilized upon macrocycle were compared with those of a closely related monomeric B/B/P FLP adduct 8, which can addition of one extra equivalent of B(C6F5)3 (BCF) to the starting material (cf. Figure 24). be stabilized upon addition of one extra equivalent of B(C6F5)3 (BCF) to the starting material (cf. Figure Figure 25a,b show the 11B MAS NMR spectra of the monomeric B/B/P FLP and the tetrameric B/P 24). FLP-CO2 macrocycle, respectively. The two boron species detected for compound 8 were assigned to boron in the FLP unit (B1, red curve) and in the BCF ligand (B2, blue curve) according to DFT calculations based on the single crystal structure. The 11B nucleus associated with B1 shows an isotropic J-coupling to 31P with 1J = 42 Hz, as determined by 2D heteronuclear J-resolved MAS NMR spectroscopy (see Figure 26) and conﬁrmed by DFT calculation. There is no J-coupling detected for B2. For compound 7, the two detected boron species originate from two slightly crystallographically diﬀerent boron positions in the tetrameric macrocycle. No isotropic J-coupling is observed (and expected by DFT calculation) in this case (data not shown).

Molecules 2019, 24, x FOR PEER REVIEW 24 of 40

species bound to phosphorus (black) and oxygen (red). Additionally, the intensity of an integral over the total superimposed spectrum is depicted (blue squares). Corresponding simulations based on the single-crystal structure are included as solid lines. The simulations consider a two-spin P–B system with a distance of 2.10 Å (black curve) and a three-spin 11B…31P…11B system with distances of 4.38 and 5.4 Å and a B…P…B angle of 117° (red curve). Furthermore, the normalized sum of both simulations is shown (blue curve). In (b), the REDOR (bottom) and the reference spectra (top) acquired for a dipolar evolution time of 0.48 ms are depicted along with corresponding line-shape simulations.

Figure 23. Experimental and simulated 31P{11B} REAPDOR curves obtained for 5. Corresponding simulations based on the single-crystal structure are included as solid lines. The simulated curve is the weighted sum of the individual curves obtained for the three possible isotopologues: a three-spin 11B…31P…11B system with distances of 2.10 and 4.38 Å (64.2%, red curve) and a B…P…B angle of 112° and two 11B…31P two-spin systems with distances of 2.10 and 4.38 Å (blue and green curves). The dashed curve is an approximate two-spin simulation, neglecting the secondary 4.38 Å distance.

3.5. Tetrameric B/P FLP–CO2 Macrocycles. The six-membered cyclic B/P FLP 6 shown in Figure 24 adds carbon dioxide under mild conditions. Upon crystallization, a unique macrocyclic tetramer 7 with bridging CO2 molecules between the individual cyclo-FLP units was obtained [85]. The spectroscopic features of this macrocycle were compared with those of a closely related monomeric B/B/P FLP adduct 8, which can be stabilized upon addition of one extra equivalent of B(C6F5)3 (BCF) to the starting material (cf. Figure 24). Molecules 2020, 25, 1400 24 of 39 Molecules 2019, 24, x FOR PEER REVIEW 25 of 40 Molecules 2019, 24, x FOR PEER REVIEW 25 of 40

Figure 24. Preparation and molecular structure of the tetrameric B/P FLP–CO2 adduct 7 and structure Figure 24. Preparation and molecular structure of the tetrameric B/P FLP–CO2 adduct 7 and structure of the monomeric B(C6F5)3 and CO2 adduct 8. of the monomeric B(C6F5)3 and CO2 adduct 8.

11 FiguresFigures 25a 25a and and 25b 25b show show the the 11 B BMAS MAS NMR NMR spectra spectra of ofthe the monomeric monomeric B/B/P B/B/P FLP FLP and and the the tetramerictetrameric B/P B/P FLP-CO FLP-CO2 2macrocycle, macrocycle, respectively. respectively. The The two two boron boron species species detected detected for forcompound compound 8 8 werewere assignedassigned to to boron boron in in the the FLP FLP unit unit (B1, (B1, red red curve) curve) and and in inthe the BCF BCF ligand ligand (B2, (B2, blue blue curve) curve) accordingaccording to to DFT DFT calculations calculations based based on on the the single single crystal crystal structure. structure. The The 11B 11nucleusB nucleus associated associated with with B1B1 shows shows an an isotropic isotropic J-coupling J-coupling to to 31 P31 Pwith with 1J =1J 42= 42 Hz, Hz, as asdetermined determined by by2D 2D heteronuclear heteronuclear J-resolved J-resolved MASMAS NMR NMR spectroscopy spectroscopy (see (see Figure Figure 26) 26) and and confirm confirmeded by by DFT DFT calculation. calculation. There There is no is J-couplingno J-coupling detecteddetected forfor B2.B2. ForFor compound compound 7 ,7 the, the two two detected detected boron boron species species originate originate from from two two slightly slightly crystallographicallycrystallographically different different boron boron positions positions in inthe the tetrameric tetrameric macrocycle. macrocycle. No No isotropic isotropic J-coupling J-coupling is is Figure 24. Preparation and molecular structure of the tetrameric B/P FLP–CO2 adduct 7 and structure observedobserved (and (and expected expected by by DFT DFT calculation) calculation) in inthis this case case (data (data not not shown). shown). of the monomeric B(C6F5)3 and CO2 adduct 8.

Figure 25. Experimental and simulated 11B MAS NMR spectra of (a) the monomeric compound 8 and FigureFigure 25. 25. ExperimentalExperimental andand simulatedsimulated 1111BB MAS MAS NMR NMR spectra spectra of of ( (aa)) the the monomeric monomeric compound compound 88 andand (b) the tetrameric FLP–CO2 adduct 7. (b(b)) the the tetrameric tetrameric FLP–CO FLP–CO22 adductadduct 7.

Figure 26. ((aa)) The The 2D 2D 1111B{B{3131P}P} heteronuclear heteronuclear J-resolved J-resolved MAS MAS NMR NMR spectrum spectrum of 8.of In8 addition. In addition to the to the positiveFigure 26.projections projections (a) The 2Dof of the11 theB{ indirect31 indirectP} heteronuclear dimension, dimension, J-resolvedthe the column column MAS extracted extracted NMR at spectrum − at2.5 ppm2.5 of ppm is 8 depicted. In is depictedaddition on the to on the the − left-handpositive projectionsside. side. A A comparison comparison of the indirect of ofthese these dimension, show show that that the the thecolumn species species extractedB1 B1shows shows at J-coupling −2.5 J-coupling ppm with is depicted with 31P nuclei,31P on nuclei, the characterizedleft-hand side. by Aan an comparison isotropic isotropic coupling coupling of these constant constant show ofthat of 42 42 theHz, Hz, specieswhile while no noB1 J-couplingJ-coupling shows J-coupling is is observed observed with for for species31 speciesP nuclei, B2. Hence,B2.characterized Hence, the 2Dthe experiment by2D an experiment isotropic can becouplingcan used be toused constant resolve to resolve bothof 42 species.Hz, both while species. This no isJ-coupling This demonstrated is demonstrated is observed in (b), showingforin (speciesb), the sumshowingB2. ofHence, rows the sumthe extracted 2Dof rows experiment at extracted21 Hz can and at −be 21 Hzused and for to species21 resolve Hz for B1 bothspecies and species. the B1 row and extractedThis the rowis demonstrated extracted at 0 Hz for at 0 B2. Hzin (b), − forshowing B2. the sum of rows extracted at −21 Hz and 21 Hz for species B1 and the row extracted at 0 Hz for B2.

Molecules 2019, 24, x FOR PEER REVIEW 26 of 40 Molecules 2020, 25, 1400 25 of 39 For both compounds, the 31P MAS NMR spectra show only a single species, even though, in the case Forof 7 both, two compounds, crystallographically the 31P MAS distinct NMR sites spectra are showexpected. only As a single illustrated species, by even Figure though, 27a, inthe the J- couplingcase of 7, twowith crystallographically the 11B nuclei (I = 3/2) distinct present sites in are 8 expected.manifests Asitself illustrated in a resolved by Figure multiplet. 27a, the Figures J-coupling 27c withand 27d the 11alsoB nuclei reveal (I a= significant3/2) present difference in 8 manifests between itself the in 31 aP resolved chemical multiplet. shift anisotropy Figure 27 parametersc,d also reveal for thea significant two compounds, difference which between was the confirmed31P chemical by shiftDFT anisotropycalculation, parameters further highlighting for the two the compounds, effect of whichintermolecular was confirmed association. by DFT calculation, further highlighting the effect of intermolecular association.

Figure 27. ExperimentalExperimental and and simulated simulated 3131P{P{1H}1H} CP/MAS CP/MAS NMR NMR spectra spectra of of8 and8 and 7 measured7 measured at at7.05 7.05 T andT and MAS MAS frequencies frequencies of of(a, (ba,) b12.5) 12.5 kHz kHz and and (c, (dc), d3.0) 3.0kHz. kHz. Note Note the thesignificant significant differences differences in the in spinningthe spinning sideband sideband patterns, patterns, indicating indicating the the potential potential of ofthe the 3131P Pchemical chemical shift shift anisotropy anisotropy for for site differentiation.differentiation. The minor component observed in 8 is attributed to anan impurityimpurity (marked byby aa cross).cross).

11 31 The molecular structures of thesethese compoundscompounds areare furtherfurther confirmedconfirmed byby 11B{B{31P} REDOR experiments (Figure 28).28). The The experimental experimental data data are found found to to be be in good agreement with simulations based on the molecular co conformationnformation obtained from the single-crystal structures. In the case of 8, the ... . significantsignificant differences differences in the the B B….P Pinternuclear internuclear distances distances for for B1 B1and and B2 B2are are easily easily detected, detected, while, while, in thein the case case of of7, 7a, adeconvoluted deconvoluted REDOR REDOR analysis analysis is is preclude precludedd by by the the close close signal signal overlap. Thus, the observed REDOR behavior corresponds to the average of both individual REDOR curves. In In this this case, case, it was again found necessary to conduct a three-spinthree-spin analysis,analysis, including the closest intermolecular ... . B….P Pdistances distances to to the the next next B/P B/ PFLP FLP molecule molecule inside inside the the macrocycle. macrocycle. Figure 29 shows complementary 31P{11B} REAPDOR data. As each of the 31P nuclei interacts with two closest 11B spins, the existence of different isotopologues must be considered, as described above. For 8, these contributions consist of a dominant three-spin component (64.2%, 2.87 and 4.32 Å, 11 ... 31 ... 11 B P B angle of 78◦, blue curve) and two minor two-spin components from the two possible 11B–31P–10B isotopologues (15.9% each), while 4% of the molecules do not contain any 11B and, thus, will yield no REAPDOR effect. Analogous simulations were made for 7, based on the intramolecular (3.29 Å) and the intermolecular closest 31P ... 11B distances (4.16 Å), as well as a 11B ... 31P ... 11B angle of 119◦. Figure 29 shows overall good agreement with the experimental data.

Molecules 2020, 25, 1400 26 of 39 Molecules 2019, 24, x FOR PEER REVIEW 27 of 40

Figure 28. 28. IndividualIndividual 11B{1131B{P}31 REDORP} REDOR (squares) (squares) and compensated and compensated REDOR REDOR (circles) (circles) curves curves(see Figure (see 11)Figure obtained 11) obtained for the forintegrated the integrated signal intensity signal intensity of (a) 8 of and (a) (8band) 7 . In (b )the7. Incase the of case 8, solid of 8, curves solid curves are two- are spintwo-spin simulations simulations based based on the on crystal the crystal structure structure with with 11B…3111BP ...distances31P distances of 2.87 of Å 2.87 (red Å curve, (red curve, B1) and B1) 4.32and Å 4.32 (blue Å (blue curve, curve, B2). B2).Black Black data datapoints points and andsolid solid curves curves show show the weighted the weighted sumsum obtained obtained if the if totalthe total signal signal intensity intensity is analyzed. is analyzed. In the In the case case of of7, 7simulations, simulations are are shown shown for for two-spin two-spin systems systems of of 11 31 intramolecularlyintramolecularly andand intermolecularlyintermolecularly interacting interacting B11B and and P31P nuclei nuclei with with internuclear internuclear distances distances of 3.29 of 3.29Å (blue Å (blue curve) curve) and 4.16and Å4.16 (red Å curve).(red curve). The blackThe bl curveack curve shows shows a three-spin a three-sp simulationin simulation based based on both on Molecules 2019, 24, x FOR31 PEER... 11 REVIEW... 31 28 of 40 bothdistances distances and aandP a 31P…11B B…31P angle of of 111° 111◦ reflectingreﬂecting the the structural structural situation situation in in the the crystal crystal structure. structure.

Figure 29 shows complementary 31P{11B} REAPDOR data. As each of the 31P nuclei interacts with two closest 11B spins, the existence of different isotopologues must be considered, as described above. For 8, these contributions consist of a dominant three-spin component (64.2%, 2.87 and 4.32 Å, 11B…31P…11B angle of 78°, blue curve) and two minor two-spin components from the two possible 11B– 31P–10B isotopologues (15.9% each), while 4% of the molecules do not contain any 11B and, thus, will yield no REAPDOR effect. Analogous simulations were made for 7, based on the intramolecular (3.29 Å) and the intermolecular closest 31P…11B distances (4.16 Å), as well as a 11B…31P…11B angle of 119°. Figure 29 shows overall good agreement with the experimental data. The aggregated character of 7 can be verified by the 31P–31P CT-DRENAR data shown in Figure 30. The monomeric FLP and the macrocyclic FLP tetramer are well differentiated based on the strength of the homonuclear 31P–31P magnetic dipole–dipole interactions. For compound 8, the closest internuclear 31P….31P distance is 8.77 Å and, thus, the dipolar recoupling effect (the difference signal 31 11 intensityFigureFigure Δ S29. ) is ExperimentalExperimental very small, andin and very simulated good agreement31PP{ {11B}B} REAPDOR REAPDOR with the curves curvessimulation for for (a (a) )8(Figure8 andand ( (bb )30a,) 77,, including including green curve). the the In 31 31 contrast,simulatedsimulated each componentscomponentsP nucleus of inof the thethe corresponding tetramercorresponding 7 is isotopologues neighbored isotopologues based by twobased on the Pon dipolar nuclei the geometriesdipolarat the distancegeometries as indicated. of 6.16as Å, leadingindicated. to a significantly stronger recoupling effect, which is again in excellent agreement with the 31 31 simulationThe aggregated (Figure 30b, character green curve). of 7 can Moreover, be verified the by ex theperimentalP– P CT-DRENAR data also differentiate data shown this in Figurematerial 30 . veryThe monomericwell from a FLP hypothetical and the macrocyclic FLP dimer, FLP for tetramer which only are well a two-spin differentiated interaction based would on the have strength to be of 31 31 considered.the homonuclear P– P magnetic dipole–dipole interactions. For compound 8, the closest internuclear 31P ... .31P distance is 8.77 Å and, thus, the dipolar recoupling effect (the difference signal intensity ∆S) is very small, in very good agreement with the simulation (Figure 30a, green curve). In contrast, each 31P nucleus in the tetramer 7 is neighbored by two 31P nuclei at the distance of 6.16 Å, leading to a significantly stronger recoupling effect, which is again in excellent agreement with the simulation (Figure 30b, green curve). Moreover, the experimental data also differentiate this material very well from a hypothetical FLP dimer, for which only a two-spin interaction would have to be considered.

Figure 30. 31P–31P CT-DRENAR curves (black squares) for (a) 8 and (b) 7 acquired for a dipolar recoupling time of 2.56 ms at 7.05 T, using a MAS frequency of 12.5 kHz with SWFTPPM [84] 1H decoupling. The simulation in (a) is based on a two-spin system with a 31P…31P distance of 8.77 Å (green curve) based on the shortest internuclear distance in the single-crystal structure. In (b), two simulations are depicted: a hypothetical dimer (two-spin system, gray curve) and the tetramer (three- spin system, green curve). Each simulation considered a 31P…31P distance of 6.16 Å and, in the latter case, a 31P…31P…31P angle of 76° based on the single-crystal structure. The agreement of the experimental data with the three-spin simulation gives clear evidence for the formation of the macrocycle.

3.6. An FLP-Cyclooctamer The six-membered cyclic B/P FLP 6 illustrated in Figure 31 shows a remarkable tendency for self-assembly at low temperatures in solution and upon crystallization, producing the cyclooctameric structure 9 with alternating chair- and boat-like conformations [86]. The boron and the phosphorus atoms are placed at the tip of the chair (B1 and P1), while, in the boat conformation, the FLP centers are part of the basis (B2 and P2) (see Figure 32). To facilitate the computation of the NMR observables, the simplified molecules 10a and 10b served as models of the corresponding cutouts of the molecular structure of 9. The 11B MAS and the 31P{1H} CP/MAS NMR spectra are shown in Figures 31c and 31d. The 11B MAS NMR spectrum features two signals at isotropic chemical shifts of −3.7 ppm and −5.4 ppm. As already shown in Figure 6, the resolution can be significantly improved by the triple- quantum single-quantum correlation experiment; the INEPT method offers an alternative (vide infra). Each of the two resonances in the 31P{1H} CP/MAS NMR spectra centered at 20.0 ppm and 17.7 ppm in Figure 31d consists of two distinct isotopologue signals (due to the effect of 31P–11B and 31P–10B J-

Molecules 2019, 24, x FOR PEER REVIEW 28 of 40

Figure 29. Experimental and simulated 31P {11B} REAPDOR curves for (a) 8 and (b) 7, including the simulated components of the corresponding isotopologues based on the dipolar geometries as Molecules 2020, 25, 1400 27 of 39 indicated.

FigureFigure 30. 3131P–P–3131P PCT-DRENAR CT-DRENAR curves curves (black (black squares) squares) for for (a ()a )88 andand (b (b) )77 acquiredacquired for for a a dipolar dipolar recouplingrecoupling time time of of 2.56 2.56 ms ms at at 7.05 7.05 T, T, using a MAS frequency of 12.5 kHzkHz with SWFTPPM [84] [84] 1HH decoupling.decoupling. The The simulation simulation in in ( a) isis basedbased onon aa two-spintwo-spin system system with with a a31 31PP...…3131PP distance distance of of 8.77 8.77 Å Å (green(green curve) curve) based based on on the shortest internuclear distance in the single-crystal structure. In In ( (bb),), two two simulationssimulations are are depicted: depicted: a hypothetical a hypothetical dimer dimer (two (two-spin-spin system, system, gray curve) gray curve) and the and tetramer the tetramer (three- spin(three-spin system, system, green curve). green curve).Each simulation Each simulation considered considered a 31P…31P a 31distanceP ... 31P of distance 6.16 Å and, of 6.16 in Åthe and, latter in 31 ... 31 ... 31 case,the latter a 31P case,…31P…31 a PP angleP of 76°P anglebased of on 76◦ thebased single-crystal on the single-crystal structure. structure. The agreement The agreement of the experimentalof the experimental data with data the with three-spin the three-spin simulation simulation gives givesclear clearevidence evidence for the for formation the formation of the of macrocycle.the macrocycle.

3.6. An FLP-Cyclooctamer 3.6. An FLP-Cyclooctamer 6 TheThe six-membered six-membered cyclic cyclic B/P B/P FLP FLP 6 illustratedillustrated in in Figure Figure 3131 showsshows a a remarkable remarkable tendency tendency for for self-assemblyself-assembly at at low low temperatures temperatures in in solution solution and and upon upon crystallization, crystallization, producing producing the the cyclooctameric cyclooctameric 9 structurestructure 9 with alternating chair- and boat-like conformations [86]. [86]. The The boron boron and and the the phosphorus phosphorus atomsatoms are placed atat thethe tiptip ofof the the chair chair (B1 (B1 and and P1), P1), while, while, in in the the boat boat conformation, conformation, the the FLP FLP centers centers are arepart part of theof the basis basis (B2 (B2 and and P2) P2) (see (see Figure Figure 32 32).). To To facilitate facilitate the the computationcomputation of the NMR observables, observables, thethe simplified simplified molecules molecules 10a10a andand 10b10b servedserved as as models models of of the the corresp correspondingonding cutouts cutouts of of the the molecular molecular structure of 9. The 11B MAS and the 31P{1H} CP/MAS NMR spectra are shown in Figure 31c,d. The 11B structure of 9. The 11B MAS and the 31P{1H} CP/MAS NMR spectra are shown in Figures 31c and 31d. MAS NMR spectrum features two signals at isotropic chemical shifts of 3.7 ppm and 5.4 ppm. The 11B MAS NMR spectrum features two signals at isotropic chemical shifts− of −3.7 ppm− and −5.4 ppm.As already As already shown shown in Figure in 6Figure, the resolution 6, the resolution can be significantly can be significantly improved improved by the triple-quantum by the triple- quantumsingle-quantum single-quantum correlation correlati experiment;on experiment; the INEPT the method INEPT off methoders an alternative offers an alternative (vide infra). ( Eachvide infra of the). 31 1 Eachtwo resonancesof the two inresonances the P{ H} in CPthe/ MAS31P{1H} NMR CP/MAS spectra NMR centered spectra at 20.0 centered ppm andat 20.0 17.7 ppm ppm and in Figure 17.7 ppm 31d 31 11 31 10 inconsists Figureof 31d two consists distinct of isotopologuetwo distinct is signalsotopologue (due sign to theals e(dueffect to of theP– effectB and of 31P–P–11B Band J-coupling). 31P–10B J- The four crystallographically distinct FLP monomers of 9 in the same conformation (chair- or boat-like) present common NMR signals [86]. Molecules 2019, 24, x FOR PEER REVIEW 29 of 40 Molecules 2019, 24, x FOR PEER REVIEW 29 of 40 coupling). The four crystallographically distinct FLP monomers of 9 in the same conformation (chair- coupling). The four crystallographically distinct FLP monomers of 9 in the same conformation (chair- orMolecules boat-like)2020, 25present, 1400 common NMR signals [86]. 28 of 39 or boat-like) present common NMR signals [86].

Figure 31. (a) Schematic representation of FLP monomer 6 and FLP-cyclooctamer 9 and (b) molecular FigureFigure 31. 31. (a(a) )Schematic Schematic representation representation of of FLP FLP monomer monomer 66 andand FLP-cyclooctamer FLP-cyclooctamer 99 andand (b (b) )molecular molecular structure of 9, as well as the respective experimental and simulated (c) 11B and (d) 31P MAS NMR structurestructure of of 99, ,as as well well as as the the respective respective experimental and and simulated ( (cc)) 11BB and and ( (dd)) 3131PP MAS MAS NMR NMR spectra of 9. spectraspectra of of 99. .

FigureFigure 32.32. Chair-Chair- and and boat-like boat-like conformational conformational geometries geometries present present in 9. To in facilitate 9. To thefacilitate DFT calculation the DFT Figure 32. Chair- and boat-like conformational geometries present in 9. To facilitate the DFT calculationof the NMR of observables, the NMR observables, the simplified the molecules simplified10a moleculesand 10b shown10a and served 10b shown as models served for theas models cutouts calculation of the NMR observables, the simplified molecules 10a and 10b shown served as models forof thethe molecularcutouts of structurethe molecular of 9, where,structure for of instance, 9, where, NMR for parametersinstance, NMR of P1 parameters and B2 were of P1 calculated and B2 for the cutouts of the molecular structure of 9, where, for instance, NMR parameters of P1 and B2 wereusing calculated10a. using 10a. were calculated using 10a. 11 TheThe assignment assignment of of experimental to to calculated 11BB NMR NMR parameters parameters and, and, therefore, therefore, to to certain The assignment of experimental to calculated 11B NMR parameters and,11 therefore, to certain conformersconformers of of the the monomeric monomeric unit is no nott straightforward, since the calculated 11BB chemical chemical shifts shifts and and conformers of the monomeric unit is not straightforward, since the calculated 11B chemical shifts and quadrupolarquadrupolar coupling parametersparameters forfor the the two two boron boron sites sites are are very very similar. similar. Key Key to theto the assignment assignment is the is quadrupolar coupling parameters31 for the two boron sites are very similar. Key to the assignment is theconsiderable considerable diff differenceerence inthe in theP 31 chemicalP chemical shift shift anisotropies anisotropies between between the the P1 P1 and and P2 P2 sites sites evident evident in the considerable difference in the 31P chemical shift anisotropies between the P1 and P2 sites evident31 inFigure Figure 33 ,33, which which is alsois also consistent consistent with with the the DFT DFT calculations. calculations. This This diff differenceerence is quite is quite evident evident in the in theP in Figure 33, which is also consistent with the DFT calculations. This difference is quite evident in the 31MASP MAS NMR NMR spinning spinning sideband sideband pattern, pattern, which which extends extend overs aover significantly a significantly wider wider spectral spectral range forrange the 31P MAS NMR spinning sideband pattern, which11 extends over a significantly wider spectral range forP1 (chair)the P1 (chair) than for than the for P2 (boat)the P2 unit.(boat) As unit. each As Beach resonance 11B resonance could becould assigned be assigned to the directlyto the directly bound for the P1 (chair) than for the P231 (boat)11 unit. As each 11B resonance could be assigned to the directly boundphosphorus phosphorus atom using atom the using 2D theP{ 2DB} CP31P{/refocused11B} CP/refocused INEPT correlation INEPT correlation experiment experiment (see Figure (see 34), boundthe boron phosphorus species at atom3.8 ppm using can the be attributed2D 31P{11B} to CP/refocused boron in a chair INEPT conformation correlation (B1) ofexperiment the monomeric (see Figure 34), the boron− species at −3.8 ppm can be attributed to boron in a chair conformation (B1) of Figureunit, while 34), the the boron nucleusspecies at in − the3.8 boat ppm conformation can be attributed (B2) gives to boron rise toin a a resonance chair conformation at about 5.3 (B1) ppm. of − The latter species experiences a slightly stronger quadrupolar coupling, as also confirmed by DFT Molecules 2019, 24, x FOR PEER REVIEW 30 of 40

Molecules 2020, 25, 1400 29 of 39 the monomeric unit, while the boron nucleus in the boat conformation (B2) gives rise to a resonance Moleculesat about 2019 ,− 245.3, x FORppm. PEER The REVIEW latter species experiences a slightly stronger quadrupolar 30coupling, of 40 as also calculations.confirmed by Obviously,DFT calculations. if the crystal Obviously, structure if the of crys thetal cyclooctamer structure of werethe cyclooctamer unknown, the were experimental unknown, the monomeric unit, while the boron nucleus in the boat conformation (B2) gives rise to a resonance datathe experimental discussed above data woulddiscussed also above allow would for alternative also allow interpretations. for alternative Forinterpretations. example, the For two example, pairs of at about −5.3 ppm. The latter species experiences a slightly stronger quadrupolar coupling, as also 11theB two and pairs31P signals of 11B and could 31P equallysignals could well beequally a result well of be a dimerica result of or a tetrameric dimeric or macrocycle,tetrameric macrocycle, or a result confirmed by DFT calculations. Obviously, if the crystal structure of the cyclooctamer were unknown, theofor experimental twoa result diﬀ oferent two data crystal different discussed structures crystal above structureswould of the also FLP ofal monomer,low the forFLP alternative monomer, which interpretations. are which mixed are inside mixed For theexample, inside powder the powder sample. 31 31 theTosample. two exclude pairs To of theexclude 11B latterand 31 theP scenario, signals latter could scenario, a Pequally CP /aBaBa wellP CP/BaBa spectrumbe a result spectrum of was a dimeric acquired, was or tetramericacquired, in which macrocycle, in cross which peaks cross arise peaks for ordirectlyarise a result for of dipolardirectly two different coupleddipolar crystal coupled nuclei. structures Withnuclei. of the theWith objective FLP the monomer, objective of selecting which of selecting are correlation mixed correlation inside peaks the powder peaks only for only the for most the sample.proximalmost proximal To exclude31P nuclei, 31 Pthe nuclei, latter this experiment scenario,this experiment a 31 wasP CP/BaBa done was withdonespectrum thewith minimum was the acquired,minimum number in numberwhich of excitation cross of excitation peaks blocks, blocks, using arisetheusing for BaBa-XY16 thedirectly BaBa-XY16 dipolar sequence. coupledsequence. Under nuclei. Under this With condition, this the condition, objective only of the only selecting P1–P2 the P1–P2correlation cross correlationcross peaks correlation only (r(P1–P2) for (ther(P1–P2) = 5.3 Å)= 5.3 is mostobservedÅ) isproximal observed (see 31P Figure (seenuclei, Figure 35this). experiment In35). contrast, In contrast, was the done dipolarthe with dipolar couplingsthe minimum couplings between number between the of excitation chemicallythe chemically blocks, equivalent equivalent31P usingspecies31P species the BaBa-XY16 are are much much sequence. weaker weaker ( rUnder(P1–P1) (r(P1–P1) this= condition, 8.4= 8.4 Å; Å;r(P2–P2) ronly(P2–P2) the= P1–P2 =10.1 10.1 cross Å Å and, and, correlationthus, thus, nono (r auto-correlation(P1–P2)auto-correlation = 5.3 peaks Å) is observed (see Figure 35). In contrast, the dipolar couplings between the chemically equivalent are seen. If the pairs pairs of of boron boron and and phosphorus phosphorus reso resonancesnances would result from two different diﬀerent crystalline 31P species are much weaker (r(P1–P1) = 8.4 Å; r(P2–P2) = 10.1 Å and, thus, no auto-correlation peaks polymorphs, no cross-correlations wouldwould bebe observed.observed. are seen. If the pairs of boron and phosphorus resonances would result from two different crystalline polymorphs, no cross-correlations would be observed.

Figure 33. Experimental and simulated slow-spinning 31P MAS NMR spectra (spinning frequency Figure 33.33. Experimental and simulated slow-spinning 31PP MAS NMR spectra (spinning(spinning frequency 3 kHz) of the cyclooctamer 9, revealing clear differences in the 31P chemical shift anisotropies of the 3 kHz) of the cyclooctamer 9, revealingrevealing clearclear didifferencesﬀerences inin thethe 3131P chemical shift anisotropies of the two phosphorus sites. Simulations of the corresponding static 31P NMR spectra are included as well. two phosphorus sites.sites. Simulation Simulationss of the corresponding static 31PP NMR NMR spectra spectra are are included included as well.

Figure 34. (a) 2D 11B{31P} CP/refocused INEPT experiment on 9. The assignment of boron species results from the covalent bond of P1 in the chair conformation of the six-membered ring with B2 in the boat conformation and vice versa for P2 and B1. (b) Horizontal slices of each species (black line) and 11B line-shape simulations (blue and red curves).

Molecules 2019, 24, x FOR PEER REVIEW 31 of 40

Figure 34. (a) 2D 11B {31P} CP/refocused INEPT experiment on 9. The assignment of boron species results from the covalent bond of P1 in the chair conformation of the six-membered ring with B2 in the boat conformation and vice versa for P2 and B1. (b) Horizontal slices of each species (black line) Molecules 2020, 25, 1400 30 of 39 and 11B line-shape simulations (blue and red curves).

31 1 Figure 35. TheThe 2D 2D 31P{P{1H}H} CP/BaBa CP/BaBa NMR NMR spectrum spectrum of of 99, ,acquired acquired with with four four BaBa BaBa blocks blocks with with four 90° 90◦ 31 pulses each each (see (see Figure Figure 10). 10). 31P Pchemical chemical shifts shifts in inthe the single-quantum single-quantum (SQ) (SQ) and and DQ DQdimensions dimensions are are assigned to the resonances revealing P1–P2 correlation peaks, whereas no P1–P1 or P2–P2 assigned to the resonances revealing P1–P2 correlation peaks, whereas no P1–P1 or P2–P2 autocorrelations are observed. autocorrelations are observed. Figure 36a shows the 11B{31P} REDOR curve of 9. Since both 11B species are strongly superimposed Figure 36a shows the 11B{31P} REDOR curve of 9. Since both 11B species are strongly in the 11B MAS NMR spectrum, only the total signal intensity can be analyzed. Alongside the superimposed in the 11B MAS NMR spectrum, only the total signal intensity can be analyzed. experimental data, three different simulated REDOR curves are depicted, which are based on the Alongside the experimental data, three different simulated REDOR curves are depicted, which are single-crystal structure obtained via X-ray diffraction. It is expected that the dephasing of the 11B based on the single-crystal structure obtained via X-ray diffraction. It is expected that the dephasing resonances is dominated by dipolar couplings to the covalently bound 31P atom of the neighboring of the 11B resonances is dominated by dipolar couplings to the covalently bound 31P atom of the FLP monomer (green curve, average internuclear B ... P distance of 2.13 Å (B ... P distances are 2.15 Å neighboring FLP monomer (green curve, average internuclear B…P distance of 2.13 Å (B…P distances (B1) and 2.11 Å (B2)). If no oligomer were formed, the strongest dipolar coupling would result from the are 2.15 Å (B1) and 2.11 Å (B2)). If no oligomer were formed, the strongest dipolar coupling would intramolecular B–P interaction, for which another simulation based on an average distance of 3.42 Å is result from the intramolecular B–P interaction, for which another simulation based on an average depicted (blue curve). A comparison with the experimental data clearly proves the formation of a distance of 3.42 Å is depicted (blue curve). A comparison with the experimental data clearly proves supramolecular structure, where strong intermolecular 11B/31P dipole–dipole interactions are present. the formation of a supramolecular structure, where strong intermolecular 11B/31P dipole–dipole Regarding this comparison, an excellent agreement between simulated and experimental data is only interactions are present. Regarding this comparison, an excellent agreement between simulated and achieved if a full three-spin system is included in the simulation (black curve). The angle between experimental data is only achieved if a full three-spin system is included in the simulation (black both dipolar vectors used for this simulation was obtained from the crystal structure (B1: 145 and B2: curve). The angle between both dipolar vectors used for this simulation was obtained from the◦ crystal 144 ). The excellent agreement between experimental data and simulation proves that considering structure◦ (B1: 145° and B2: 144°). The excellent agreement between experimental data and simulation the interaction with the two closest spins is sufficient to quantitatively reproduce the REDOR data in proves that considering the interaction with the two closest spins is sufficient to quantitatively this case. The analogous 31P{11B} REAPDOR data, shown in Figure 36b, produce the same conclusion. reproduce the REDOR data in this case. The analogous 31P{11B} REAPDOR data, shown in Figure 36b As discussed above, the experimental REAPDOR curve is the sum of the three distinct contributions , produce the same conclusion. As discussed above, the experimental REAPDOR curve is the sum of arising from the boron isotopologues in the expected 64.2:15.9:15.9 ratio. the three distinct contributions arising from the boron isotopologues in the expected 64.2:15.9:15.9 ratio.

Molecules 2020, 25, 1400 31 of 39 Molecules 2019, 24, x FOR PEER REVIEW 32 of 40

11 31 FigureFigure 36. 36. ((aa)) Experimental Experimental and and simulated simulated 11B{B{31P}P} REDOR REDOR curves curves of of 99. .Two-spin Two-spin simulations simulations based based onon the the shortest shortest intermolecular intermolecular distance, distance, including including a a calculated calculated J-J-anisotropyanisotropy of of 89 89 Hz Hz (green (green curve) curve) andand the the intramolecular intramolecular distance distance (b (bluelue curve). curve). Black Black curve: curve: Three-sp Three-spinin simulation simulation based based on on the the two two closestclosest boron–phosphorus boron–phosphorus distancesdistances inin thethe specific specific dipolar dipolar geometry geometry present present in in the the crystal crystal structure. structure. (b ) 31 11 (b)P{ 31P{B}11 REAPDORB} REAPDOR curves curves for 9 showingfor 9 showing simulations simulations for the three for distinctthe three isotopologue distinct isotopologue contributions contributionsto the overall to response the overall (black) response curve. (black) curve. Finally, the 31P–31P CT-DRENAR experiment can be used to confirm the macrocyclic structure Finally, the 31P–31P CT-DRENAR experiment can be used to confirm the macrocyclic structure of of 9. In analogy to the tetrameric compound 7, only three-spin simulations based on the actual 9. In analogy to the tetrameric compound 7, only three-spin simulations based on the actual crystal crystal structure of 9 are suitable to represent the experimental data, directly excluding mono- and structure of 9 are suitable to represent the experimental data, directly excluding mono- and dimeric dimeric structures. Interestingly, P1 shows a much stronger dephasing than P2 despite the identical structures. Interestingly, P1 shows a much stronger dephasing than P2 despite the identical internuclear 31P ... 31P distances of 5.3 Å. This can be explained by the strongly differing 31P chemical internuclear 31P…31P distances of 5.3 Å. This can be explained by the strongly differing 31P chemical shift anisotropies evident in Figure 33, which are known to influence DRENAR experiments [72–75]. shift anisotropies evident in Figure 33, which are known to influence DRENAR experiments [72–75]. For the octameric structures, it is important to determine if this influence originates from the 31P For the octameric structures, it is important to determine if this influence originates from the 31P chemical shift anisotropy (CSA) of the coupling partners or the observed nucleus itself. As can be chemical shift anisotropy (CSA) of the coupling partners or the observed nucleus itself. As can be deduced from the simulations in Figure 37c, an increased 31P CSA of the coupling partners strongly deduced from the simulations in Figure 37c, an increased 31P CSA of the coupling partners strongly decreases the maximum DRENAR difference signal, while corresponding changes in the 31P CSA decreases the maximum DRENAR difference signal, while corresponding changes in the 31P CSA of of the observed nucleus show no influence within the regime of experimentally determined ∆σ the observed nucleus show no influence within the regime of experimentally determined Δσ up to up to 124 ppm. Therefore, P2 interacting with two P1 phosphorus spins with a large 31P CSA of 124 ppm. Therefore, P2 interacting with two P1 phosphorus spins with a large 31P CSA of 124 ppm 124 ppm shows a weaker DRENAR effect compared to P1 interacting with two P2 phosphorus spins shows a weaker DRENAR effect compared to P1 interacting with two P2 phosphorus spins (CSA = (CSA = 78 ppm). To increase the overall DRENAR effect, longer dipolar mixing times can be applied 78 ppm). To increase the overall DRENAR effect, longer dipolar mixing times can be applied for for weakly interacting 31P nuclei, as demonstrated in Figure 37b. Again, a good agreement between weakly interacting 31P nuclei, as demonstrated in Figure 37b. Again, a good agreement between three- three-spin simulations and experimental data confirms the findings discussed. In conclusion, the spin simulations and experimental data confirms the findings discussed. In conclusion, the CT- CT-DRENAR experiments are well suited to determine 31P ... 31P internuclear distances and spin DRENAR experiments are well suited to determine 31P…31P internuclear distances and spin system system geometries, as long as CSA effects are taken into account in the simulation. This requires that geometries, as long as CSA effects are taken into account in the simulation. This requires that the the latter can be determined accurately from slow-spinning MAS NMR spectra. latter can be determined accurately from slow-spinning MAS NMR spectra.

Molecules 2020, 25, 1400 32 of 39 Molecules 2019, 24, x FOR PEER REVIEW 33 of 40

FigureFigure 37.37. 3131P CT-DRENARCT-DRENAR curves curves for for species species P1 (blackP1 (black squares) squares) and P2 and (blue P2 squares) (blue squares) of the octameric of the octamericcompound compound9 acquired 9 using acquired a dipolar using mixing a dipolar time mixing of (a, ctime, d) 1.28 of ( msa, c, and d) 1.28 (b) 2.56 ms msand at (b 7.05) 2.56 T withms at a 7.05MAS T spinningwith a MAS frequency spinning of 12.5 freq kHzuency and of SWFTPPM 12.5 kHz1 Hand decoupling. SWFTPPM In 1 (Ha) anddecoupling. (b), simulations In (a) and for each(b), simulationsspecies based for on each a two-spin species (dashed based on curves) a two-spin and a three-spin (dashed curves) system (solidand a curves) three-spin are included system (solid based curves)on a 31P are... 31includedP distance based of 5.30on a Å.31P In…31 addition,P distance experimentally of 5.30 Å. In addition, determined experimentally31P CSAs were determined considered 31P CSAsin all simulations,were considered as well in asall the simulations, angles between as we dipole–dipolell as the angles and between CSA interaction dipole–dipole tensors and from CSA the interactionsingle-crystal tensors structure. from Thethe single-crystal agreement of structure. the experimental The agreement data with ofthe thethree-spin experimental simulations data with gives the three-spinevidence for simulations the formation gives of evidence the octameric for the macrocycle. formation Additionally, of the octameric simulations macrocycle. for P1 Additionally, using various 31 31 simulationsP CSAs for for (c )P1 the using coupling various nuclei 31P CSAs and (d for) the (c) observed the coupling nucleus nuclei are and depicted (d) the to observed demonstrate nucleus the areP 31 31 depictedCSA inﬂuence to demonstrate on the P– the P31P CT-DRENAR CSA influence curve. on the 31P–31P CT-DRENAR curve. 4. Conclusions 4. Conclusions Advanced solid-state NMR experiments can provide key information regarding the structure and Advanced solid-state NMR experiments can provide key information regarding the structure bonding character of aggregated FLPs and FLP adducts. If aggregation proceeds via intermolecular and bonding character of aggregated FLPs and FLP adducts. If aggregation proceeds via association of the borane/phosphane units, 11B chemical shift and nuclear electric quadrupolar coupling intermolecular association of the borane/phosphane units, 11B chemical shift and nuclear electric constants, as well as isotropic 1J spin–spin coupling constants, are particularly useful in characterizing quadrupolar coupling constants, as well as isotropic 1J spin–spin coupling constants, are particularly the strength of such intermolecular associations. Table1 gives a comprehensive summary of all the useful in characterizing the strength of such intermolecular associations. Table 1 gives a relevant NMR observables obtained in this ﬁeld. comprehensive summary of all the relevant NMR observables obtained in this field.

Table 1. Solid-state NMR parameters of FLP Aggregates 1–9 a.

11B Parameters 31P Parameters J-Coupling

𝐢𝐬𝐨 31 Com 𝒊𝒔𝒐 Coord. 𝜹𝐂𝐒 ± J( P- 𝜹 ± CQ ± 0.05 ηQ Δσ ± 5 ησ Calc. ΔJ p. Species 𝑪𝑺 31P 0.5 11B) 0.5 (ppm) (MHz) ± 0.05 (ppm) ± 0.05 (Hz) species (ppm) ± 5 (Hz) −4.6 1.21 0.63 48.9 93.8 0.83 49 1 B P −4.2 1.34 0.62 57.0 120.1 0.74 41/22 −2.1 0.84 0.48 11.7 49.5 0.43 2 B P −4.6 0.94 0.42 5.8 55.4 0.40 13/17

Molecules 2020, 25, 1400 33 of 39

Table 1. Solid-state NMR parameters of FLP Aggregates 1–9 a.

11B Parameters 31P Parameters J-Coupling

Comp. iso Coord. iso 31 11 δ CQ ηQ δ ∆σ ησ J( P B) Calc. ∆J Species CS ± ± 31P CS ± ± − 0.5 (ppm) 0.05 (MHz) 0.05 0.5 (ppm) 5 (ppm) 0.05 5 (Hz) (Hz) ± species ± ± 4.6 1.21 0.63 48.9 93.8 0.83 49 1 B − P 4.2 1.34 0.62 57.0 120.1 0.74 41/22 − 2.1 0.84 0.48 11.7 49.5 0.43 2 B − P 4.6 0.94 0.42 5.8 55.4 0.40 13/17 − 14.6 1.69 0.35 17.4 99 0.57 3 B − P 17.6 1.67 0.36 26.4 61.8 92 − 9.5 1.10 0.20 21.8 66 B1 − P1 11.3 1.11 0.25 26.6 62 − 8.2 2.30 0.30 4 B2, B4 6.7/5.8 2.32/2.33 0.30 9.3 1.00 0.20 20.8 70 B3 − P2 11.3 1.03 0.26 22.3 70 − 9.5 1.08 0.30 14.8 80 B1 − P1 10.9 1.17 0.11 18.0 69 92 − 7.3 2.38 0.25 B2 7.5 2.49 0.26 8.3 1.05 0.30 13.2 82 B3 − P2 12.0 1.04 0.17 14.1 78 92 − 6.5 2.37 0.28 B4 5 5.92 2.27 0.33 8.9 1.10 0.30 11.7 78 B5 − P3 13.0 1.13 0.06 12.7 75 92 − 6.9 2.36 0.28 B6 6.13 2.24 0.33 8.9 1.09 0.20 10.1 70 B7 − P4 7.5 2.21 0.15 B8

1.3 2.17 0.50 B1A 14.6 90 0.42 1.0 2.30 0.52 7 P 1.2 2.04 0.47 B1B 15.5 101 0.41 13 1.2 2.18 0.52 − 5.8 2.69 0.24 B1 3.8 2.78 0.17 8 0.7 2.13 0.05 3.9 43 0.95 42 B2 P 0.8 2.23 0.10 2.8 56 0.38 41 − − B1 3.7 1.70 0.64 P2 20.0 78.4 0.96 38 − 5.0 1.79 0.57 24.5 70.0 0.86 33 89 9 − B2 5.4 1.79 0.62 P1 17.7 123.8 0.68 40 − 4.0 1.90 0.72 24.7 128.0 0.74 28 89 − a DFT calculated values are highlighted in gray. Distinct sites are identiﬁed according to their crystallographic information.

Likewise, REDOR and REAPDOR offer the possibility of distance measurements. For the analysis of such experiments on FLP (adduct) aggregates, two-spin models are usually no longer appropriate, and the full distance geometry of the heteronuclear spin system must generally be taken into account. In many cases, the approximate analysis in terms of three-spin systems involving the shortest distances tends to be an acceptable approximation. Detailed systematic studies on model compounds indicate that the sensitivity of REDOR and REAPDOR methods is limited to maximum distances of about 6 Å; measurements toward longer distances in FLP-related compounds are impeded by spin–spin relaxation effects [79]. As shown in the example of the tetrameric CO2 adduct 7 and the octameric FLP 9, the evaluation of homonuclear 31P–31P dipole–dipole interaction strengths via DQ-DRENAR, and other homonuclear recoupling experiments are particularly useful for differentiating between different Molecules 2020, 25, 1400 34 of 39 models of intermolecular associations. While, in the present study, all the compounds discussed could be characterized independently by single-crystal X-ray diffraction, the NMR methodologies and database developed here will be of great value for learning about the structural organization of poly- and nano-crystalline materials or surface FLP deposits to be studied in future endeavors, including, for instance, the development of heterogeneous FLP catalysts. Initial efforts in this direction are already appearing in the literature. For example, a heterogeneous H2 splitting catalyst based on an intermolecular FLP was designed by reacting an organoboron Lewis acid covalently attached to a silica support with the Lewis base P(t-Bu)3 [87] and characterized by standard MAS NMR methods. Introducing more advanced NMR methodologies to such systems will certainly be of great benefit in characterizing their more intricate structural features. Ultimately, the structure/function correlations discovered along these lines will be useful for optimizing their catalytic performance.

5. Materials and Methods Solid-state MAS NMR studies were conducted on a Bruker Avance III 300 spectrometer, and DSX 400 and 500 MHz spectrometers (magnetic ﬂux density of 7.05, 9.39, and 11.74 T, respectively). Typical MAS frequencies were 12.5 kHz in 4-mm NMR double- and triple-resonance probes. 1H decoupling was typically used applying the SWFTPPM decoupling scheme [84] with 60 kHz nutation frequency. 31P{1H} CP/MAS NMR spectra were typically acquired via ramped (90%–100% of maximum Hartmann–Hahn nutation frequency of 43 kHz) cross-polarization [88] at contact times of 500 µs and typical relaxation delays of 15 s. All the two-dimensional J-resolved, correlation, and dipolar recoupling experiments were measured at MAS frequencies of 10–15 kHz and nutation frequencies near 50 kHz. Stationary magnetization conditions were ensured by saturation combs preceding the pulse sequences. The ﬁrst and second J-evolution periods in the INEPT block were optimized for maximum signal intensity giving delays of 4.3 and 5.0 ms, far below the value suggested by 1/(2J) to avoid signal loss due to transverse relaxation. Spectral deconvolution was done with the DMFIT software [39] (version 2011). DFT calculations of NMR parameters were conducted based on crystallographic input or geometry-optimized molecular models such as those shown in Figure 30 (TURBOMOLE (version 6.5 and 7.1) [43,89] combined with the TPSS functional [90] and an Ahlrich’s def2-TZVP basis set [91]. Chemical shift calculations also used TURBOMOLE and the def2-TZVP basis set in combination with the B3LYP functional [92,93]. Quadrupolar coupling parameters were calculated on a GGA DFT level using GAUSSIAN (version GAUSSIAN09) [42] and the B97-D functional [94]. The def2-TZVP basis set obtained from the EMSL database [95,96] could be used with additional functions for boron from the cc-pCVTZ basis set to enhance the accuracy near the nucleus. Anisotropic J-coupling constants were calculated with the software CPL implemented in the ADF program [97–99]. The calculation was performed using the PBE0 hybrid functional [100] and JCPL basis sets [101], which are based on the TZ2P ZORA basis set, but which also include additional functions with high exponents for an improved representation of the electron density close to the nuclei.

Author Contributions: This review was written by H.E. and R.K., based on experimental work conducted by R.K., M.B., and T.W. under the supervision of H.E. and M.R.H. The compounds studied here were synthesized in the laboratory of G.E. under his and G.K.’s supervision. Funding for this work was awarded to G.E. and H.E. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by Deutsche Forschungsgemeinschaft, grant number SFB858, subproject A1 (G.E. and H.E.). R.K. and T.W. thank the Fonds der Chemischen Industrie for doctoral stipends. Acknowledgments: All the crystallographic characterization of the materials discussed here was done by C.G. Daniuluc, Organisch-Chemisches Institut, WWU Münster. Conﬂicts of Interest: The authors declare no conﬂict of interest. Molecules 2020, 25, 1400 35 of 39

References

1. Wiegand, T.; Eckert, H.; Grimme, S. Frustrated Lewis Pairs I: Uncovering and Understanding; Erker, G., Stephan, D.W., Eds.; Topics in Current Chemistry; Springer-Verlag: Berlin, Germany, 2013; ISBN 978-3-642-36696-3. 2. Welch, G.C.; Juan, R.R.S.; Masuda, J.D.; Stephan, D.W. Reversible, Metal-Free Hydrogen Activation. Science 2006, 314, 1124–1126. [CrossRef][PubMed] 3. Spies, P.; Schwendemann, S.; Lange, S.; Kehr, G.; Fröhlich, R.; Erker, G. Metal-Free Catalytic Hydrogenation of Enamines, Imines, and Conjugated Phosphinoalkenylboranes. Angew. Chem. Int. Ed. 2008, 47, 7543–7546. [CrossRef][PubMed] 4. Spies, P.; Kehr, G.; Bergander, K.; Wibbeling, B.; Fröhlich, R.; Erker, G. Metal-free Dihydrogen Activation Chemistry: Structural and Dynamic Features of Intramolecular P/B pairs. Dalton Trans. 2009, 9, 1534–1541. [CrossRef] 5. Stephan, D.W. “Frustrated Lewis pair” Hydrogenations. Org. Biomol. Chem. 2012, 10, 5740–5746. [CrossRef] 6. Er˝os,G.; Nagy, K.; Mehdi, H.; Pápai, I.; Nagy, P.; Király, P.; Tárkányi, G.; Soós, T. Catalytic Hydrogenation with Frustrated Lewis Pairs: Selectivity Achieved by Size-Exclusion Design of Lewis Acids. Chem. Eur. J. 2012, 18, 574–585. [CrossRef] 7. Stephan, D.W.; Erker, G. Frustrated Lewis Pairs: Metal-free Hydrogen Activation and More. Angew. Chem. Int. Ed. 2010, 49, 46–76. [CrossRef] 8. Mömming, C.M.; Otten, E.; Kehr, G.; Fröhlich, R.; Grimme, S.; Stephan, D.W.; Erker, G. Reversible Metal-Free Carbon Dioxide Binding by Frustrated Lewis Pairs. Angew. Chem. Int. Ed. 2009, 48, 6643–6646. [CrossRef] 9. Sajid, M.; Klose, A.; Birkmann, B.; Liang, L.; Schirmer, B.; Wiegand, T.; Eckert, H.J.; Lough, A.; Fröhlich, R.;

G. Daniliuc, C.; et al. Reactions of Phosphorus/Boron Frustrated Lewis Pairs with SO2. Chem. Sci. 2013, 4, 213–219. [CrossRef] 10. Sajid, M.; Stute, A.; Cardenas, A.J.P.; Culotta, B.J.; Hepperle, J.A.M.; Warren, T.H.; Schirmer, B.; Grimme, S.; Studer, A.; Daniliuc, C.G.; et al. N,N-Addition of Frustrated Lewis Pairs to Nitric Oxide: An Easy Entry to a Unique Family of Aminoxyl Radicals. J. Am. Chem. Soc. 2012, 134, 10156–10168. [CrossRef] 11. Sajid, M.; Elmer, L.-M.; Rosorius, C.; Daniliuc, C.G.; Grimme, S.; Kehr, G.; Erker, G. Facile Carbon Monoxide Reduction at Intramolecular Frustrated Phosphane/Borane Lewis Pair Templates. Angew. Chem. Int. Ed. 2013, 52, 2243–2246. [CrossRef] 12. Ekkert, O.; Miera, G.G.; Wiegand, T.; Eckert, H.; Schirmer, B.; Petersen, J.L.; Daniliuc, C.G.; Fröhlich, R.; Grimme, S.; Kehr, G.; et al. Remarkable Coordination Behavior of Alkyl Isocyanides Toward Unsaturated Vicinal Frustrated P/B Lewis Pairs. Chem. Sci. 2013, 4, 2657–2664. [CrossRef]

13. Ménard, G.; Stephan, D.W. H2 Activation and Hydride Transfer to Oleﬁns by Al(C6F5)3-Based Frustrated Lewis Pairs. Angew. Chem. Int. Ed. 2012, 51, 8272–8275. [CrossRef][PubMed] 14. Dureen, M.A.; Brown, C.C.; Stephan, D.W. Deprotonation and Addition Reactions of Frustrated Lewis Pairs with Alkynes. Organometallics 2010, 29, 6594–6607. [CrossRef] 15. Jiang, C.; Blacque, O.; Berke, H. Activation of Terminal Alkynes by Frustrated Lewis Pairs. Organometallics 2010, 29, 125–133. [CrossRef] 16. Ullrich, M.; Seto, K.S.-H.; Lough, A.J.; Stephan, D.W. 1,4-Addition Reactions of Frustrated Lewis Pairs to 1,3-Dienes. Chem. Commun. 2009, 2335–2337. [CrossRef] 17. Xu, B.-H.; Yanez, R.A.A.; Nakatsuka, H.; Kitamura, M.; Fröhlich, R.; Kehr, G.; Erker, G. Reaction of Frustrated Lewis Pairs with Ketones and Esters. Chem. Asian J. 2012, 7, 1347–1356. [CrossRef] 18. Wiegand, T.; Eckert, H.; Grimme, S. Solid-State NMR as a Spectroscopic Tool for Characterizing Phosphane–Borane Frustrated Lewis Pairs. In Frustrated Lewis Pairs I: Uncovering and Understanding; Topics in Current Chemistry; Erker, G., Stephan, D.W., Eds.; Springer: Berlin, Germany, 2013; pp. 291–345. ISBN 978-3-642-36697-0. 19. Wiegand, T.; Siedow, M.; Eckert, H.; Kehr, G.; Erker, G. Structural Characterization of Frustrated Lewis Pairs and Their Reaction Products Using Modern Solid-State NMR Spectroscopy Techniques. Isr. J. Chem. 2015, 55, 150–178. [CrossRef] 20. Wiegand, T.; Eckert, H.; Ekkert, O.; Fröhlich, R.; Kehr, G.; Erker, G.; Grimme, S. New Insights into Frustrated Lewis Pairs: Structural Investigations of Intramolecular Phosphane–Borane Adducts by Using Modern Solid-State NMR Techniques and DFT Calculations. J. Am. Chem. Soc. 2012, 134, 4236–4249. [CrossRef] Molecules 2020, 25, 1400 36 of 39

21. Elmer, L.-M.; Kehr, G.; Daniliuc, C.G.; Siedow, M.; Eckert, H.; Tesch, M.; Studer, A.; Williams, K.; Warren, T.H.; Erker, G. The Chemistry of a Non-Interacting Vicinal Frustrated Phosphane/Borane Lewis Pair. Chem. Eur. J. 2017, 23, 6056–6068. [CrossRef] 22. Sajid, M.; Kehr, G.; Wiegand, T.; Eckert, H.; Schwickert, C.; Pöttgen, R.; Cardenas, A.J.P.; Warren, T.H.; Fröhlich, R.; Daniliuc, C.G.; et al. Noninteracting, Vicinal Frustrated P/B-Lewis Pair at the Norbornane Framework: Synthesis, Characterization, and Reactions. J. Am. Chem. Soc. 2013, 135, 8882–8895. [CrossRef] 23. Liedtke, R.; Scheidt, F.; Ren, J.; Schirmer, B.; Cardenas, A.J.P.; Daniliuc, C.G.; Eckert, H.; Warren, T.H.; Grimme, S.; Kehr, G.; et al. Frustrated Lewis Pair Modification by 1,1-Carboboration: Disclosure of a Phosphine Oxide Triggered Nitrogen Monoxide Addition to an Intramolecular P/B Frustrated Lewis Pair. J. Am. Chem. Soc. 2014, 136, 9014–9027. [CrossRef][PubMed] 24. Ge, F.; Türkyilmaz, F.; Daniliuc, C.G.; Siedow, M.; Eckert, H.; Kehr, G.; Erker, G. A 1,1-Carboboration Route to Bora-Nazarov Systems. Chem. Asian J. 2015, 10, 2497–2502. [CrossRef][PubMed] 25. Dong, S.; Wang, L.; Wang, T.; Daniliuc, C.G.; Brinkkötter, M.; Eckert, H.; Kehr, G.; Erker, G. Formation and Reactions of Active Five-membered Phosphane/Borane Frustrated Lewis Pair Ring Systems. Dalton Trans. 2018, 47, 4449–4454. [CrossRef][PubMed] 26. Apperley, D.C.; Harris, R.K.; Hodgkinson, P. Solid-State NMR: Basic Principles and Practice; Momentum Press: New York, NY, USA, 2012; ISBN 978-1-60650-352-2. 27. Abragam, P.A. Principles of Nuclear Magnetism; Oxford University Press: New York, NY, USA, 1961; ISBN 978-0-19-852014-6. 28. Keeler, J. Understanding NMR Spectroscopy; John Wiley & Sons: Chichester, UK, 2011; ISBN 978-1-119-96493-3. 29. Duer, M.J. Solid State NMR Spectroscopy: Principles and Applications; Wiley-Blackwell: Malden, MA, USA, 2002; ISBN 978-0-632-05351-3. 30. Slichter, C.P. Principles of Magnetic Resonance, 3rd ed.; Springer Series in Solid-State Sciences; Springer-Verlag: Berlin, Germany, 1990; ISBN 978-3-540-50157-2. 31. Pell, A.J.; Pintacuda, G.; Grey, C.P. Paramagnetic NMR in Solution and the Solid State. Prog. Nucl. Magn. Reson. Spectrosc. 2019, 1, 111. [CrossRef] 32. Harris, R.K.; Becker, E.D.; Cabral De Menezes, S.M.; Granger, P.; Hoffman, R.E.; Zilm, K.W. Further Conventions for NMR Shielding and Chemical Shifts IUPAC Recommendations 2008. Solid State Nucl. Magn. Reson. 2008, 33, 41–56. [CrossRef] 33. Andrew, E.R.; Bradbury, A.; Eades, R.G. Nuclear Magnetic Resonance Spectra from a Crystal Rotated at High Speed. Nature 1958, 182, 1659. [CrossRef] 34. Kaupp, M.; Bühl, M.; Malkin, V.G. Calculation of NMR and EPR Parameters: Theory and Applications; John Wiley & Sons: Weinheim, Germany, 2006; ISBN 978-3-527-60496-8. 35. Herzfeld, J.; Berger, A.E. Sideband Intensities in NMR Spectra of Samples Spinning at the Magic Angle. J. Chem. Phys. 1980, 73, 6021–6030. [CrossRef] 36. Lin, Y.-L.; Cheng, Y.-S.; Ho, C.-I.; Guo, Z.-H.; Huang, S.-J.; Org, M.-L.; Oss, A.; Samoson, A.; Chan, J.C.C. Preparation of Fibril Nuclei of Beta-Amyloid Peptides in Reverse Micelles. Chem. Commun. 2018, 54, 10459–10462. [CrossRef] 37. Penzel, S.; Oss, A.; Org, M.-L.; Samoson, A.; Böckmann, A.; Ernst, M.; Meier, B.H. Spinning Faster: Protein NMR at MAS Frequencies up to 126 kHz. J. Biomol. NMR 2019, 73, 19–29. [CrossRef] 38. Man, P.P. Quadrupolar Interactions. In Encyclopedia of Nuclear Magnetic Resonance; John Wiley & Sons: Chichester, UK, 2007; pp. 3838–3848. ISBN 978-0-471-49082-1. 39. Massiot, D.; Fayon, F.; Capron, M.; King, I.; Calvé, S.L.; Alonso, B.; Durand, J.-O.; Bujoli, B.; Gan, Z.; Hoatson, G. Modelling One- and Ttwo-Dimensional Solid-State NMR Spectra. Magn. Reson. Chem. 2002, 40, 70–76. [CrossRef] 40. Bak, M.; Rasmussen, J.T.; Nielsen, N.C. SIMPSON: A General Simulation Program for Solid-State NMR Spectroscopy. J. Magn. Reson. 2011, 213, 366–400. [CrossRef][PubMed] 41. Perras, F.A.; Widdifield, C.M.; Bryce, D.L. QUEST—QUadrupolar Exact SofTware: A fast Graphical Program for the Exact Simulation of NMR and NQR Spectra for Quadrupolar Nuclei. Solid State Nucl. Magn. Reson. 2012, 45–46, 36–44. [CrossRef][PubMed] 42. Frisch, M.; Trucks, G.; Schlegel, H.; Scuseria, G.; Robb, M.; Cheeseman, J.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G.; et al. Gaussian 09, Revision B.01; Gaussian, Inc.: Wallingford, CT, USA, 2009. Available online: https://gaussian.com (accessed on 18 March 2020). Molecules 2020, 25, 1400 37 of 39

43. TURBOMOLE v6.5 2013, a Development of University of Karlsruhe and Forschungszentrum Karlsruhe GmbH, 1989–2007, TURBOMOLE GmbH, since 2007. Available online: http://www.turbomole.com (accessed on 18 March 2020). 44. Blaha, P.; Schwarz, K.; Madsen, G.K.H.; Kvasnicka, D.; Luitz, J. WIEN2k, An Augmented Plane Wave Plus Local Orbitals Program for Calculating Crystal Properties, 2nd ed.; Vienna University of Technology: Vienna, Austria, 2001; ISBN 3-9501031-1. 45. Segall, M.D.; Lindan, P.J.D.; Probert, M.J.; Pickard, C.J.; Hasnip, P.J.; Clark, S.J.; Payne, M.C. First-Principles Simulation: Ideas, Illustrations and the CASTEP Code. J. Phys. Cond. Matter 2002, 14, 2717–2744. [CrossRef] 46. Medek, A.; Harwood, J.S.; Frydman, L. Multiple-Quantum Magic-Angle Spinning NMR: A New Method for the Study of Quadrupolar Nuclei in Solids. J. Am. Chem. Soc. 1995, 117, 12779–12787. [CrossRef] 47. Amoureux, J.-P.; Fernandez, C.; Frydman, L. Optimized Multiple-Quantum Magic-Angle Spinning NMR Experiments on Half-Integer Quadrupoles. Chem. Phys. Lett. 1996, 259, 347–355. [CrossRef] 48. Amoureux, J.-P.; Fernandez, C.; Steuernagel, S. Z-Filtering in MQMAS NMR. J. Magn. Reson. A 1996, 123, 116–118. [CrossRef] 49. Vaara, J.; Jokisaari, J.; Wasylishen, R.E.; Bryce, D.L. Spin-Spin Coupling Tensors as Determined by Experiment and Computational Chemistry. Prog. Nucl. Magn. Reson. Spectrosc. 2002, 41, 233–304. [CrossRef] 50. Bryce, D.L.; Wasylishen, R.E. Indirect Nuclear Spin-Spin Coupling Tensors in Diatomic Molecules: A Comparison of Results Obtained by Experiment and First Principles Calculations. J. Am. Chem. Soc. 2000, 122, 3197–3205. [CrossRef] 51. Olivieri, A.C. Quadrupolar Eﬀects in the CPMAS NMR Spectra of Spin-1/2 Nuclei. J. Magn. Reson. 1989, 81, 201–205. 52. Harris, R.K.; Olivieri, A.C. Quadrupolar Eﬀects Transferred to Spin-1/2 Magic-Angle Spinning Spectra of Solids. Prog. Nucl. Magn. Reson. Spectrosc. 1992, 24, 435–456. [CrossRef] 53. Hierso, J.-C. Indirect Nonbonded Nuclear Spin–Spin Coupling: A Guide for the Recognition and Understanding of “Through-Space” NMR J Constants in Small Organic, Organometallic, and Coordination Compounds. Chem. Rev. 2014, 114, 4838–4867. [CrossRef] 54. Wiegand, T.; Eckert, H.; Ren, J.; Brunklaus, G.; Fröhlich, R.; Daniliuc, C.G.; Lübbe, G.; Bussmann, K.; Kehr, G.; Erker, G.; et al. Indirect “No-Bond” 31P 31P Spin–Spin Couplings in P,P-[3]Ferrocenophanes: Insights from ··· Solid-State NMR Spectroscopy and DFT Calculations. J. Phys. Chem. A 2014, 118, 2316–2331. [CrossRef] 55. Zilm, K.W.; Grant, D.M. High-Resolution NMR Spectra with J-Couplings in Solids. J. Magn. Reson. 1982, 48, 524–526. [CrossRef] 56. Fyfe, C.A.; Wong-Moon, K.C.; Huang, Y.; Grondey, H. INEPT Experiments in Solid-State NMR. J. Am. Chem. Soc. 1995, 117, 10397–10398. [CrossRef] 57. Franke, D.; Hudalla, C.; Eckert, H. Heteronuclear X-Y Double Quantum MAS NMR in Crystalline Inorganic Solids. Applications for Indirect Detection and Spectral Editing of Rare-Spin Resonances. Solid State Nucl. Magn. Reson. 1992, 1, 33–40. [CrossRef] 58. Cadars, S.; Sein, J.; Duma, L.; Lesage, A.; Pham, T.N.; Baltisberger, J.H.; Brown, S.P.; Emsley, L. The Refocused INADEQUATE MAS NMR Experiment in Multiple Spin-Systems: Interpreting Observed Correlation Peaks and Optimising Lineshapes. J. Magn. Reson. 2007, 188, 24–34. [CrossRef] 59. Gullion, T.; Schaefer, J. Rotational-Echo Double-Resonance NMR. J. Magn. Reson. 1989, 81, 196–200. [CrossRef] 60. Gullion, T.; Vega, A.J. Measuring Heteronuclear Dipolar Couplings for I = 1/2, S > 1/2 Spin Pairs by REDOR and REAPDOR NMR. Prog. Nucl. Magn. Reson. Spectrosc. 2005, 47, 123–136. [CrossRef] 61. Gullion, T. Rotational-Echo, Double-Resonance NMR. In Modern Magnetic Resonance; Webb, G.A., Ed.; Springer Science & Business Media: Dordrecht, The Netherlands, 2007; ISBN 978-1-4020-3910-2. 62. Chan, J.C.C.; Eckert, H. Dipolar Coupling Information in Multispin Systems: Application of a Compensated REDOR NMR Approach to Inorganic Phosphates. J. Magn. Reson. 2000, 147, 170–178. [CrossRef] 63. Gullion, T. Measurement of Dipolar Interactions between Spin-1/2 and Quadrupolar Nuclei by Rotational-Echo, Adiabatic-Passage, Double-Resonance NMR. Chem. Phys. Lett. 1995, 246, 325–330. [CrossRef] 64. Gullion, T. Detecting 13C–17O Dipolar Interactions by Rotational-Echo, Adiabatic-Passage, Double-Resonance NMR. J. Magn. Reson. 1995, 117, 326–329. [CrossRef] Molecules 2020, 25, 1400 38 of 39

65. Chopin, L.; Vega, S.; Gullion, T. A MAS NMR Method for Measuring 13C 17O Distances. J. Am. Chem. Soc. − 1998, 120, 4406–4409. [CrossRef] 66. Nimerovsky, E.; Goldbourt, A. Distance Measurements between Boron and Carbon at Natural Abundance Using Magic Angle Spinning REAPDOR NMR and a Universal Curve. Phys. Chem. Chem. Phys. 2012, 14, 13437–13443. [CrossRef] 67. Feike, M.; Graf, R.; Schnell, I.; Jäger, C.; Spiess, H.W. Structure of Crystalline Phosphates from 31P Double-Quantum NMR Spectroscopy. J. Am. Chem. Soc. 1996, 118, 9631–9634. [CrossRef] 68. Saalwächter, K.; Lange, F.; Matyjaszewski, K.; Huang, C.-F.; Graf, R. BaBa-xy16: Robust and Broadband Homonuclear DQ Recoupling for Applications in Rigid and Soft Solids up to the Highest MAS Frequencies. J. Magn. Reson. 2011, 212, 204–215. [CrossRef] 69. Geen, H.; Gottwald, J.; Graf, R.; Schnell, I.; Spiess, H.W.; Titman, J.J. Elucidation of Dipolar Coupling Networks under Magic-Angle Spinning. J. Magn. Reson. 1997, 125, 224–227. [CrossRef] 70. Hohwy, M.; Jakobsen, H.J.; Edén, M.; Levitt, M.H.; Nielsen, N.C. Broadband Dipolar Recoupling in the Nuclear Magnetic Resonance of Rotating Solids: A compensated C7 Pulse Sequence. J. Chem. Phys. 1998, 108, 2686–2694. [CrossRef] 71. Lee, Y.K.; Kurur, N.D.; Helmle, M.; Johannessen, O.G.; Nielsen, N.C.; Levitt, M.H. Eﬃcient Dipolar Recoupling in the NMR of Rotating Solids. A Sevenfold Symmetric Radiofrequency Pulse Sequence. Chem. Phys. Lett. 1995, 242, 304–309. [CrossRef] 72. Ren, J.; Eckert, H. A Homonuclear Rotational Echo Double-Resonance Method for Measuring Site-Resolved Distance Distributions in I=1/2 Spin Pairs, Clusters, and Multispin Systems. Angew. Chem. Int. Ed. 2012, 51, 12888–12891. [CrossRef] 73. Ren, J.; Eckert, H. DQ-DRENAR: A New NMR Technique to Measure Site-Resolved Magnetic Dipole-Dipole Interactions in Multispin-1/2 Systems: Theory and Validation on Crystalline Phosphates. J. Chem. Phys. 2013, 138, 164201. [CrossRef] 74. Ren, J.; Eckert, H. Measurement of Homonuclear Magnetic Dipole–Dipole Interactions in Multiple Spin-1/2 Systems Using Constant-Time DQ-DRENAR NMR. J. Magn. Reson. 2015, 260, 46–53. [CrossRef] 75. Ren, J.; Eckert, H. DQ-DRENAR with Back-to-Back (BABA) Excitation: Measuring Homonuclear Dipole-Dipole Interactions in Multiple Spin 1/2 Systems. Solid State Nucl. Magn. Reson. 2015, 71, 11–18. [CrossRef] 76. Pines, A.; Gibby, M.G.; Waugh, J.S. Proton-Enhanced Nuclear Induction Spectroscopy. A Method for High Resolution NMR of Dilute Spins in Solids. J. Chem. Phys. 1972, 56, 1776–1777. [CrossRef] 77. Hartmann, S.R.; Hahn, E.L. Nuclear Double Resonance in the Rotating Frame. Phys. Rev. 1962, 128, 2042–2053. [CrossRef] 78. Wiegand, T. Structure, Reactivity, and Cooperativity of Functional Organophosphor Compounds: Development and Application of Modern Solid-State NMR and DFT Strategies. PhD Thesis, University of Münster, Münster, Germany, 7 January 2013. 79. Siedow, M. Charakterisierung Frustrierter Lewis-Paar-Addukte mittels Kombination moderner Festkörper-NMR-Methoden mit quantenchemischen Berechnungen. PhD Thesis, University of Münster, Münster, Germany, 18 April 2018. 80. Knitsch, R. Structural Characterization of Novel Organoboron Compounds by Application of Magnetic Resonance Spectroscopy and DFT Calculations. PhD Thesis, University of Münster, Münster, Germany, 24 April 2019. 81. Erdmann, M.; Wiegand, T.; Blumenberg, J.; Eckert, H.; Ren, J.; Daniliuc, C.G.; Kehr, G.; Erker, G. Formation, Structural Characterization, and Reactions of a Unique Cyclotrimeric Vicinal Lewis Pair Containing

(C6F5)2P-Lewis Base and (C6F5)BH-Lewis Acid Components. Dalton Trans. 2014, 43, 15159–15169. [CrossRef] 82. Sajid, M.; Lawzer, A.; Dong, W.; Rosorius, C.; Sander, W.; Schirmer, B.; Grimme, S.; Daniliuc, C.G.; Kehr, G.; Erker, G. Carbonylation Reactions of Intramolecular Vicinal Frustrated Phosphane/Borane Lewis Pairs. J. Am. Chem. Soc. 2013, 135, 18567–18574. [CrossRef] 83. Wang, L.; Dong, S.; Daniliuc, C.G.; Liu, L.; Grimme, S.; Knitsch, R.; Eckert, H.; Hansen, M.R.; Kehr, G.; Erker, G. Formation of Macrocyclic Ring Systems by Carbonylation of Trifunctional P/B/B Frustrated Lewis Pairs. Chem. Sci. 2018, 9, 1544–1550. [CrossRef] Molecules 2020, 25, 1400 39 of 39

84. Chandran, C.V.; Madhu, P.K.; Kurur, N.D.; Bräuniger, T. Swept-Frequency Two-Pulse Phase Modulation (SWf-TPPM) Sequences with Linear Sweep Profile for Heteronuclear Decoupling in Solid-State NMR. Magn. Reson. Chem. 2008, 46, 943–947. [CrossRef] 85. Jie, X.; Sun, Q.; Daniliuc, C.G.; Knitsch, R.; Hansen, M.R.; Eckert, H.; Kehr, G.; Erker, G. Cycloaddition Reactions of an Active Cyclic Phosphane/Borane Pair with Alkenes, Alkynes, and Carbon Dioxide. Chem. Eur. J. 2020, 26, 1269–1273. [CrossRef] 86. Jie, X.; Daniliuc, C.G.; Knitsch, R.; Hansen, M.R.; Eckert, H.; Ehlert, S.; Grimme, S.; Kehr, G.; Erker, G. Aggregation Behavior of a Six-Membered Cyclic Frustrated Phosphane/Borane Lewis Pair: Formation of a Supramolecular Cyclooctameric Macrocyclic Ring System. Angew. Chem. Int. Ed. 2019, 58, 882–886. [CrossRef] 87. Xing, J.Y.; Buffet, J.C.; Rees, N.H.; Norby, P.; O’Hare, D. Hydrogen Cleavage by Solid-Phase Frustrated Lewis Pairs. Chem. Commun. 2016, 52, 10478–10481. [CrossRef] 88. Peersen, O.B.; Wu, X.L.; Kustanovich, I.; Smith, S.O. Variable-Amplitude Cross-Polarization MAS NMR. J. Magn. Reson. A 1993, 104, 334–339. [CrossRef] 89. Ahlrichs, R.; Bär, M.; Häser, M.; Horn, H.; Kölmel, C. Electronic Structure Calculations on Workstation Computers: The Program System Turbomole. Chem. Phys. Lett. 1989, 162, 165–169. [CrossRef] 90. Tao, J.; Perdew, J.P.; Staroverov, V.N.; Scuseria, G.E. Climbing the Density Functional Ladder: Nonempirical Meta–Generalized Gradient Approximation Designed for Molecules and Solids. Phys. Rev. Lett. 2003, 91, 146401. [CrossRef] 91. Weigend, F.; Ahlrichs, R. Balanced Basis Sets of Split Valence, Triple Zeta Valence and Quadruple Zeta Valence Quality for H to Rn: Design and Assessment of Accuracy. Phys. Chem. Chem. Phys. 2005, 7, 3297–3305. [CrossRef] 92. Becke, A.D. Density-Functional Thermochemistry. III. The Role of Exact Exchange. J. Chem. Phys. 1993, 98, 5648–5652. [CrossRef] 93. Stephens, P.J.; Devlin, F.J.; Chabalowski, C.F.; Frisch, M.J. Ab Initio Calculation of Vibrational Absorption and Circular Dichroism Spectra Using Density Functional Force Fields. J. Phys. Chem. 1994, 98, 11623–11627. [CrossRef] 94. Grimme, S. Semiempirical Hybrid Density Functional with Perturbative Second-Order Correlation. J. Chem. Phys. 2006, 124, 034108. [CrossRef] 95. Feller, D. The Role of Databases in Support of Computational Chemistry Calculations. J. Comput. Chem. 1996, 17, 1571–1586. [CrossRef] 96. Schuchardt, K.L.; Didier, B.T.; Elsethagen, T.; Sun, L.; Gurumoorthi, V.; Chase, J.; Li, J.; Windus, T.L. Basis Set Exchange: A Community Database for Computational Sciences. J. Chem. Inf. Model. 2007, 47, 1045–1052. [CrossRef] 97. Baerends, E.J.; Ziegler, T.; Atkins, A.J.; Autschbach, J.; Bashford, D.; Baseggio, O.; Berces, A.; Bickelhaupt, F.M.; Bo, C.; Boerritger, P.M.; et al. ADF 2017, Theoretical Chemistry, Vrije Universiteit, Amsterdam, documentation. Available online: http://www.scm.com (accessed on 18 March 2020). 98. Autschbach, J.; Ziegler, T. Nuclear Spin-Spin Coupling Constants from Regular Approximate Relativistic Density Functional Calculations. I. Formalism and Scalar Relativistic Results for Heavy Metal Compounds. J. Chem. Phys. 2000, 113, 936–947. [CrossRef] 99. Autschbach, J.; Ziegler, T. Nuclear Spin-Spin Coupling Constants from Regular Approximate Relativistic Density Functional Calculations. II. Spin-Orbit Coupling Effects and Anisotropies. J. Chem. Phys. 2000, 113, 9410–9418. [CrossRef] 100. Adamo, C.; Barone, V. Towards Reliable Density Functional Methods without Adjustable Parameters: The PBE0 Model. J. Chem. Phys. 1999, 110, 6158–6170. [CrossRef] 101. Bryce, D.L.; Autschbach, J. Relativistic Hybrid Density Functional Calculations of Indirect Nuclear Spin-Spin Coupling Tensors–Comparison with Experiment for Diatomic Alkali Metal Halides. Can. J. Chem. 2009, 87, 927–941. [CrossRef]

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).