Exploiting the Electronic Tuneability of Carboranes as Supports for Frustrated Lewis Pairs

Abstract

The first example of a carborane with a catecholborolyl substituent, [1-Bcat-2-Ph-closo-1,2-C₂B₁₀H₁₀] (1), has been prepared and characterized and shown to act as the Lewis acid component of an intermolecular frustrated Lewis pair in catalyzing a Michael addition. In combination with B(C₆F₅)₃ the C-carboranylphosphine [1-PPh₂-closo-1,2-C₂B₁₀H₁₁] (IVa) is found to be comparable with PPh₂(C₆F₅) in its ability to catalyze hydrosilylation, whilst the more strongly basic B-carboranylphosphine [9-PPh₂-closo-1,7-C₂B₁₀H₁₁] (V) is less effective and the very weakly basic species [μ-2,2′-PPh-{1-(1′-1′,2′-closo-C₂B₁₀H₁₀)-1,2-closo-C₂B₁₀H₁₀}] (IX) is completely ineffective. Base strengths are rank-ordered via measurement of the ¹J ³¹P-⁷⁷Se coupling constants of the phosphineselenides [1-SePPh₂-closo-1,2-C₂B₁₀H₁₁] (2), [9-SePPh₂-closo-1,7-C₂B₁₀H₁₁] (3), and [SePPh₂(C₆F₅)] (4).

1. Introduction

The recognition by Stephan and co-workers, little more than a decade ago [1], that H₂ could be reversibly activated using sterically-encumbered main group Lewis acid (LA)/Lewis base (LB) pairs has given rise to the burgeoning field of frustrated Lewis pair (FLP) chemistry [2,3,4,5,6,7,8]. FLPs can co-exist on the same molecule (intramolecular FLPs) or be on different molecules (intermolecular FLPs). Since its inception, the breadth of FLP chemistry has expanded considerably, and now, as summarized in a recent review [9], impacts upon small-molecule activation, organic chemistry, radical chemistry, transition-metal chemistry, enzyme models, polymers and materials, and surface chemistry.

Carborane chemistry is a well-established and wide-ranging field with thousands of derivatives known and a huge number of diverse applications now established for carborane-containing species [10]. Thus far, however, FLP chemistry and carborane chemistry have not intersected, in spite of the fact that the carborane scaffold offers a number of unique advantages for potential FLPs including high chemical and thermal stability, the ability to act as an electron-donating or electron-accepting support dependent on the vertex substituted (with no significant difference in the steric bulk of the carborane) [11,12,13], and further tuneability of electronic (and steric) FLP properties through isomerization, cage derivatization, or substitution [10].

We now report the first examples of intermolecular carborane-supported FLP chemistry, through (i) the synthesis of a catecholborolyl (Bcat) carborane (the LA) and its catalytic activation of a Michael addition reaction in combination with PPh₃, and (ii) a comparison of C- and B-carboranylphosphines (the LB) in combination with B(C₆F₅)₃ to catalyze a hydrosilylation reaction, demonstrating the unique advantage of carborane scaffolds in tuning the strength of FLP components.

2. Results and Discussion

2.1. Synthesis and Characterization of Compound 1

Following deprotonation of [1-Ph-closo-1,2-C₂B₁₀H₁₁] in tetrahydrofuran (THF) by ⁿBuLi and subsequent exchange of solvent for toluene, 2-Br-1,3,2-benzodioxaborole (BcatBr) in toluene was added and the reaction mixture heated to reflux overnight. The solvent was removed and the product extracted into petroleum ether. Unreacted phenylcarborane was removed via vacuum sublimation, leaving the product [1-Bcat-2-Ph-closo-1,2-C₂B₁₀H₁₀] (1) in a 45% isolated yield. As far as we are aware, compound 1 is the first example of a carborane with a catecholborolyl substituent. C-substituted pinacolborolyl (Bpin) carboranes are known (see, e.g., References [14,15]) (there is one recent report of a B-substituted pinacolborolyl carborane [16]) and several C-substituted diazaborolyl carboranes have been prepared, e.g., References [17,18].

Compound 1 is a moderately air-sensitive, but significantly moisture-sensitive, colorless solid, initially characterized using elemental analysis, mass spectrometry, and by ¹H and ¹¹B nuclear magnetic resonance (NMR) spectroscopies. The latter features the resonance due to the catecholborolyl B atom at δ 28.8 ppm in C₆D₆, easily identified by its high frequency chemical shift and lack of ¹H coupling. The carboranyl boron resonances in the ¹¹B{¹H} NMR spectrum appear with relative integrals 1:1:2:2:2:2 from high frequency to low frequency suggest a molecule with time-averaged C_s molecular symmetry in solution.

Ultimately compound 1 was unambiguously characterized crystallographically. Single crystals suitable for an X-ray diffraction study were grown by cooling a solution of 1 in C₆H₅F and a perspective view of a single molecule is shown in Figure 1. The C_s symmetry in solution was not retained in the solid state as the Ph and Bcat units were twisted by ≈19° in a conrotatory manner with respect to the least-squares plane through atoms B100C1C2C21. The C1−C2 distance in 1, 1.6840(15) Å, sits within the range of such distances in related compounds (see, e.g., References [17,18]), whilst the C1−B100 distance, 1.5703(16) Å, was comparable to that (1.565(4) Å) in [1-Bpin-2,3,4,5,6,7,8,9,10,11-Me₁₀-12-(4′-Br-C₆H₄)-closo-1-CB₁₁]⁻ [14] but significantly shorter than that (1.6046(19) Å) in [1-B(OMe)₂-2-ⁱPr-closo-1,2-C₂B₁₀H₁₀] [15], the only other compounds with {BO₂} fragments bonded to a carborane C atom to have been structurally characterized.

2.2. Carborane-Supported Components of Intermolecular FLPs to Catalyze Michael Addition

The steric bulk and electron-withdrawing nature of the phenylcarborane fragment results in B100 being both sterically-encumbered and highly Lewis acidic, and consequently compound 1 is an ideal potential LA component of an intermolecular FLP. To investigate this, we have studied the involvement of 1 as co-catalyst in the Michael addition of dimethyl malonate (I) to 3-buten-2-one (II) to afford dimethyl 2-(3-oxobutyl) malonate (III), a classic Michael addition reaction (Scheme 1). Such reactions are known to be catalyzed by phosphines alone [19], but enhanced rates have been observed if a species with the potential to act as a Lewis acid is also present because the LA and LB form an FLP (see, e.g., References [20,21,22]).

In the presence of 10 mol% PPh₃ in CD₃CN at room temperature, a 1:1 mixture of I and II affords III in 43% yield after 6 h and 64% yield after 24 h (entries 1 and 2,

Table 1. Results for catalysis of the Michael addition reaction in Scheme 1 by PPh₃, the FLP 1/PPh₃ and V. ¹

Entry	Catalyst(s)	Time (h)	Yield (%) 2
1	PPh3	6	43
2	PPh3	24	64
3	1/PPh33	6	56
4	1/PPh3	24	76
5	V	6	85
6	V	24	92

¹ Conditions: J. Young NMR tube; 1:1 molar ratio of I and II; 10 mol% of catalyst(s); CD₃CN solution; room temperature. ² Yield (average of two runs) determined by ¹H NMR spectroscopy relative to mesitylene internal standard. ³ It was established via NMR spectroscopy that 1 and PPh₃ did not form an adduct.

). No reaction was observed in the absence of a catalyst nor in the presence of only the Bcat carborane 1. When, however, an intermolecular FLP of 10 mol% 1 and 10 mol% PPh₃ was used as a catalyst, the yield of III was 56% after 6 h and 76% after 24 h (entries 3 and 4). This demonstrates the co-operative nature of the two components, with the enhancement in catalysis fully consistent with them acting as a frustrated Lewis pair and represents the first time that a carborane-containing species has been used as a component of an FLP. The results are at least as good as those obtained for the same reaction using either PhBpin/PPh₃ as an intermolecular FLP or 1-Bpin-2-PPh₂-C₆H₄ as an intramolecular FLP [22].

An interesting alternative to using PPh₃ as a catalyst for this Michael addition reaction would be to have the LB functionality on a carborane cage. However, the C-carboranylphosphines [1-PR₂-closo-1,2-C₂B₁₀H₁₁] [R = Ph (IVa), ⁱPr (IVb), ^tBu (IVc) (Scheme 2)] were all inactive in catalyzing the reaction, presumably a consequence of their relatively low basicity because of the strong electron-acceptor property of the carborane when substituted at C. In contrast, it is well established that a carborane substituted at B is electron-releasing [11,12,13], and we therefore prepared the B-substituted carboranylphosphine [9-PPh₂-closo-1,7-C₂B₁₀H₁₁] (V) [13] and tested it as the single catalyst for Michael addition, finding it to be significantly more effective than PPh₃ (Table 1, entries 5 and 6). Note that IVa and V are related as simple positional isomers. Replacing a Ph group in PPh₃ with a C-bound carborane cage (affording IVa) reduces the basicity of the phosphine and switches off the Michael addition reaction, whilst replacing a Ph in PPh₃ with a B-bound carborane cage (affording V) enhances this particular catalysis. This clearly demonstrates the potential of electronically-flexible carborane scaffolds in controlling the properties of substituents and so optimizing catalysis.

2.3. Carborane-Supported Components of Intermolecular FLPs to Catalyze Hydrosilylation

Although the C-substituted carboranylphosphines IV are too weakly basic for Michael addition, they can be, as part of an FLP, highly effective in the hydrosilylation of dimethylfulvene (Scheme 3). In combination with B(C₆F₅)₃, the weak Lewis base PPh₂(C₆F₅) has been shown by Paradies and co-workers to catalyze this reaction effectively, whilst the strong Lewis base P^tBu₃ has no activity (it is assumed that this is due to the formation of the silylium salt [^tBu₃P-SiPh₂Me][HB(C₆F₅)₃]) [23]. This is therefore an ideal reaction in which to study the effect of controlling the basicity of the Lewis base component of the FLP and we have investigated the effectiveness of carborane supports in affording that control.

In

Table 2. Results for catalysis of the hydrosilylation reaction in Scheme 2 using the FLPs B(C₆F₅)₃/PPh₂(C₆F₅), B(C₆F₅)₃/IVa, B(C₆F₅)₃/V, and B(C₆F₅)₃/IX. ¹

Entry	Catalyst(s)	Time (min)	Yield (%) 2
1	B(C6F5)3/PPh2(C6F5)	11	89
2	B(C6F5)3/IVa 3	12	88
3	B(C6F5)3/V 3	26	80
4	B(C6F5)3/IX 3	-	0 4

¹ Conditions: J. Young NMR tube; 1:1 molar ratio of LA and LB components of FLP (10 mol% of each); CD₂Cl₂ solution; room temperature. ² Yield (average of two runs) determined by ¹H NMR spectroscopy relative to mesitylene internal standard. ³ It was established by NMR spectroscopy that the LA and LB components did not form an adduct. ⁴ The reagents immediately turned deep-red, indicative of the formation of oligomerized products (see Reference [23]).

, the intermolecular FLPs formed by B(C₆F₅)₃ with PPh₂(C₆F₅), IVa and V are compared for their efficiency in catalyzing this reaction. The FLP B(C₆F₅)₃/IVa was fully comparable with B(C₆F₅)₃/PPh₂(C₆F₅) in efficiency, both combinations producing nearly 90% product yield after only 11–12 min (entries 1 and 2), whilst the FLP from B(C₆F₅)₃ with the more strongly basic V was inferior to both, affording the product only 80% yield after more than twice the time (entry 3). We also investigated the effect on the reaction of using the FLP formed from B(C₆F₅)₃ and the very weakly basic bis(carboranyl)phosphine [μ-2,2′-PPh-{1-(1′-1′,2′-closo-C₂B₁₀H₁₀)-1,2-closo-C₂B₁₀H₁₀}] (IX) [24] (Scheme 2). In this case significant amounts of oligomerized product were immediately observed (entry 4), consistent with the results obtained by Paradies when no base was used [23]. These results clearly establish that this reaction was very sensitive to the strength of the Lewis base component; if the base is too strong (e.g., P^tBu₃) there is no catalysis [23], whilst if it is too weak (e.g., IX), the base plays no part in the chemistry and the Lewis acid catalyzes oligomerization. Between these two extremes, the base acts as an FLP with the B(C₆F₅)₃ Lewis acid, which catalyzes hydrosilylation, with weaker bases performing somewhat better.

2.4. Synthesis and Characterization of Phosphineselenides 2, 3 and 4

In an attempt to understand the relative efficiencies of the phosphines PPh₃, IVa, V, IX, and PPh₂(C₆F₅) as stand-alone Lewis bases or as components of FLPs, we have attempted to rank their basicities via formation of the appropriate selenide. This is because it is well established that, in the absence of significant intra- or intermolecular H-bonding contacts, pK_b of phosphines correlates almost linearly with the magnitude of the ¹J ³¹P-⁷⁷Se coupling constants of the corresponding selenide (see, e.g., References [25,26].

The selenides [1-SePPh₂-closo-1,2-C₂B₁₀H₁₁] (2, derived from IVa), [9-SePPh₂-closo-1,7-C₂B₁₀H₁₁] (3, derived from V) and [SePPh₂(C₆F₅)] (4) were prepared in good yields using the straightforward procedure of heating to reflux the appropriate phosphine and excess Se in toluene. All three pale or colorless compounds were crystalline solids that were initially characterized using elemental analysis, mass spectrometry, and ¹H and ³¹P{¹H} NMR spectroscopies, plus ¹¹B{¹H} (for 2 and 3), ⁷⁷Se (for 3), and ¹⁹F{¹H} (for 4) NMR studies.

Compounds 2, 3, and 4 were also studied crystallographically, and perspective views of single molecules together with key molecular parameters are presented in Figure 2, Figure 3, and Figure 4, respectively. In 2, there appears to be a preferred orientation of the {SePPh₂} fragment relative to the carborane cage with the torsion angle C2−C1−P−Se only 5.73(7)°, allowing the Se atom and the relatively protonic H bound to C2 to approach to within 2.752(17) Å, substantially less than the sum of their van der Waals radii of ≈3.10 Å [27].

In all cases, the ³¹P{¹H} NMR shifts of the selenides 2, 3, and 4 were observed at significantly higher frequency than those of the precursor phosphine (δ 50.7 ppm vs 25.3 ppm for 2, +3.8 ppm vs −48.2 ppm for 3, +20.4 ppm vs −25.0 ppm for 4, all comparisons made in the same solvent). Most importantly, ⁷⁷Se satellites reveal ¹J ³¹P-⁷⁷Se coupling constants of 799 Hz for 2, 704 Hz for 3 (confirmed using the ⁷⁷Se NMR spectrum), and 774 Hz for 4. SePPh₃ (¹J ³¹P-⁷⁷Se = 732 Hz) [25] and SeIX (¹J ³¹P-⁷⁷Se = 891 Hz) [24] are known species.

Based on the magnitude of these ¹J_PSe values, the ranking of the base strengths of the phosphines was V (most basic) > PPh₃ > PPh₂(C₆F₅) > IVa > SeIX (least basic). This is fully consistent with the observation that V acts as the best stand-alone phosphine for catalyzing the Michael addition reaction but is the worst LB component of an FLP with B(C₆F₅)₃ in catalyzing the hydrosilylation reaction, known to be favored by less basic phosphines [23]. Comparison of the coupling constants of 2 and 3 dramatically illustrates the different basicities of carboranylphosphines substituted at C (compound IVa, weakly basic) versus those substituted at B (compound V, strongly basic). In an alternative description, starting from PPh₃, notional replacement of one Ph by (C₆F₅) reduces the basicity (as expected), whilst notional replacement of Ph by a C-carborane cage, forming [1-PPh₂-closo-1,2-C₂B₁₀H₁₁] (IVa), reduced the basicity even further (C-substituted carborane cage is more electron-withdrawing than Ph). On the other hand, notional replacement of one Ph in PPh₃ by a B-carborane cage, affording [9-PPh₂-closo-1,7-C₂B₁₀H₁₁] (V), increased the basicity (B-substituted carborane cage is less electron-withdrawing than Ph).

These results confirm that the nature of the site of substitution of a carborane significantly affects the strength of an appended Lewis base and, by extension, an appended Lewis acid. In principle, the acid and/or base strength can be further tuned by varying the size of the carborane, its structural type (closo, neutral; nido, anionic), its isomeric form, the nature of any heteroatoms present (e.g., metal fragment vertices), and the nature of additional substituents at B and/or at C. Thus, carborane scaffolds have the unique potential to offer an exceptional degree of control over the Lewis acid and/or Lewis base strength of appended groups, which is likely to be highly important in constructing useful FLP combinations for a variety of catalytic applications. Future contributions will explore further these possibilities.

3. Experimental Section

3.1. General Considerations

All experiments were performed, unless otherwise stated, under an atmosphere of dry nitrogen using standard Schlenk or glovebox techniques, with some subsequent manipulations and purifications carried out in air. Tetrahydrofuran (THF) was distilled from sodium/benzophenone, dichloromethane (DCM) from CaH₂, and toluene and petroleum ether (40–60 °C, petrol) from sodium. All solvents were freeze-pump-thawed three times prior to use. Deuterated solvents were stored over 4 Å molecular sieves. [1-PPh₂-closo-1,2-C₂B₁₀H₁₁] (IVa) [28], [1-PⁱPr₂-closo-1,2-C₂B₁₀H₁₁] (IVb) [29], [1-P^tBu₂-closo-1,2-C₂B₁₀H₁₁] (IVc) [30], [9-PPh₂-closo-1,7-C₂B₁₀H₁₁] (V) [13], [μ-2,2′-PPh-{1-(1′-1′,2′-closo-C₂B₁₀H₁₀)-1,2-closo-C₂B₁₀H₁₀}] (IX) [24], and [1-Ph-closo-1,2-C₂B₁₀H₁₁] [31,32] were prepared according to the literature. All other reagents were purchased from Sigma Aldrich Ltd. (Gillingham, UK), Fluorochem Ltd. (Hadfield, UK) or Alfa Aesar (Heysham, UK) and used without further purification. NMR spectra were recorded at 298 K using a Bruker AVIII-400 spectrometer (Bruker BioSpin AG, Fallenden, Switzerland), with chemical shifts reported relative to the residual protonated solvent peaks (¹H) or to external standards (¹¹B; BF₃∙OEt₂). Elemental analyses were conducted using an Exeter CE-440 elemental analyser (Exeter Analytical Inc., North Chelmsford, MA, USA). Electron ionization mass spectrometry (EIMS) was carried out on a Bruker Microtof II mass spectrometer (Bruker Daltonik GmbH, Bremen, Germany) at the University of Edinburgh.

3.1.1. Synthesis and Characterization of [1-Bcat-2-Ph-closo-1,2-C₂B₁₀H₁₀] (1)

[1-Ph-closo-1,2-C₂B₁₀H₁₁] (300 mg, 1.36 mmol) was dried under vacuum and dissolved in anhydrous THF (20 mL). The solution was cooled to 0 °C before ⁿBuLi (1.02 mL, 1.6M, 1.63 mmol) was added dropwise. The colorless solution turned pale pink and was stirred at 0 °C for 0.5 h before being warmed to room temperature, then heated to 65°C for 1 h. The pale-yellow solution was allowed to cool to room temperature and concentrated to dryness. Anhydrous toluene (25 mL) was added. The Schlenk tube was covered in foil and the pale-yellow solution cooled to −78 °C before the addition of a toluene solution of 2-bromo-1,3,2-benzodioxaborole (324 mg, 1.63 mmol), resulting in a purple solution and a blue precipitate. The mixture was heated to reflux overnight. The purple solution was transferred via cannula to a second Schlenk tube along with anhydrous toluene washings (2 × 20 mL) and concentrated to a purple solid. This was extracted with anhydrous petrol (2 × 50 mL) and the soluble materials evaporated to reveal a white solid. Excess phenyl carborane was removed via vacuum sublimation leaving the product [1-Bcat-2-Ph-closo-1,2-C₂B₁₀H₁₀] (1) (205 mg, 0.61 mmol, 45%). C₁₄H₁₉B₁₁O₂ requires C 49.7, H 5.66; found C 49.2, H 5.66%. ¹H NMR (400 MHz, C₆D₆): δ 7.44–7.41 (m, 2H, C₆H₅), 6.74–6.71 (m, 1H, C₆H₅), 6.67–6.61 (m, 4H), 6.50–6.53 (m, 2H). ¹¹B{¹H}[¹¹B] NMR (128 MHz, C₆D₆): δ 29.0 (s, 1B, Bcat), 2.4 (s [d, J_BH = 150 Hz], 1B), −2.4 (s [d, J_BH = 150 Hz], 1B), −7.3 (2B), −8.3 (2B), −9.7 (2B), −10.7 (2B). EIMS: envelope centered on m/z 338.2 (M⁺).

3.1.2. General Synthesis and Characterization of Phosphine Selenides 2, 3, and 4

The phosphine (typically 0.25–0.5 mmol) was dissolved in toluene (typically 10–15 mL) and ≈10 equivalents of elemental selenium were added. Under N₂, the reagents were heated to reflux overnight. Excess Se was filtered off and washed with DCM. The filtrate and washings were evaporated to afford essentially pure colorless or pale-colored solids.

[1-SePPh₂-closo-1,2-C₂B₁₀H₁₁] (2) Colorless, 56% yield. C₁₄H₂₁B₁₀PSe requires C 41.3, H 5.20; found C 40.8, H 5.06%. ¹H NMR (400 MHz, CDCl₃): δ 8.28–8.22 (m, 4H, C₆H₅), 7.64–7.52 (m, 6H, C₆H₅), 4.80 (br. s, 1H, C_cageH). ¹¹B{¹H} NMR (128 MHz, CDCl₃): δ −0.5 (1B), −2.5 (1B), −7.2 (2B), −11.0 (2B), −12.8 (4B). ³¹P{¹H} NMR (162.0 MHz, CDCl₃): δ 50.66 (s + Se satellites, ¹J_PSe = 799 Hz). EIMS: envelope centered on m/z 407.1 (M⁺).

[9-SePPh₂-closo-1,7-C₂B₁₀H₁₁] (3) Pale yellow, 71% yield. C₁₄H₂₁B₁₀PSe requires C 41.3, H 5.20; found C 42.2, H 5.35%. ¹H NMR (400 MHz, CD₂Cl₂): δ 8.04–7.98 (m, 4H, C₆H₅), 7.47–7.43 (m, 6H, C₆H₅), 3.13 (br. s, 2H, C_cageH). ¹¹B{¹H} NMR (128 MHz, C₆D₆): δ −3.4 (1B), −4.7 (1B), −5.9 (2B), −9.9 (1B), −11.9 (1B), −12.7 (2B), −16.3 (2B). ³¹P{¹H} NMR (162.0 MHz, C₆D₆): δ 3.76 (q [¹J_PB = 164 Hz] + Se satellites, ¹J_PSe = 704 Hz). ⁷⁷Se NMR (76.4 MHz, C₆D₆): δ −230.54 (d, ¹J_PSe = 704 Hz). EIMS: envelope centered on m/z 408.2 (M⁺).

[SePPh₂(C₆F₅)] (4) Pale pink, 57% yield. C₁₈H₁₀F₅PSe requires C 50.1, H 2.34; found C 50.1, H 2.34%. ¹H NMR (400 MHz, CDCl₃): δ 7.99–7.92 (m, 4H, C₆H₅), 7.58–7.47 (m, 6H, C₆H₅). ¹⁹F NMR (376.5 MHz, CDCl₃): δ −126.9 (s, 2F, C₆F₅), −147.1 (s, 2F, C₆F₅), −158.8 (s, 1F, C₆F₅). ³¹P{¹H} NMR (162.0 MHz, CDCl₃): δ 20.42 (s + Se satellites, ¹J_PSe = 774 Hz). EIMS: envelope centred on m/z 431.9 (M⁺).

3.2. Crystallographic Studies

Compound 1, crystal data: C₁₄H₁₉B₁₁O₂, M = 338.20, monoclinic, P2₁/c, a = 15.0049(6), b = 6.78992(18), c = 19.0043(6) Å, β = 112.135(4)°, U = 1793.49(11) ų, Z = 4, D_c = 1.253 Mg m⁻³, μ = 0.069 mm⁻¹, F(000) = 696. 37112 data to θ_max = 31.23° were collected at 120.01(10) K on a Rigaku Oxford Diffraction SuperNova diffractometer using Mo-K_α X-radiation. A total of 5441 unique reflections (R_int = 0.0424) were used to solve (using SHELXS [33]) and refine (using SHELXL [34]) the structure within the Olex2 [35] package. R₁ = 0.0480, wR₂ = 0.1189 for data with I ≥ 2σ(I), S (all data) = 1.073, E_max, E_min = 0.38, −0.23 eÅ⁻³, respectively. CCDC 1848620.

Compound 2, crystal data: C₁₄H₂₁B₁₀PSe, M = 407.34, monoclinic, P2₁/c, a = 8.7241(4), b = 25.2149(11), c = 9.4588(4) Å, β = 111.615(2)°, U = 1934.41(15) ų, Z = 4, D_c = 1.399 Mg m⁻³, μ = 2.018 mm⁻¹, F(000) = 816. 58543 data to θ_max = 34.07° were collected at 100.00(10) K on a Bruker X8 APEXII diffractometer using Mo-K_α X-radiation. A total of 7897 unique reflections (R_int = 0.0451) were used to solve and refine the structure, as for compound 1. R₁ = 0.0256, wR₂ = 0.0592 for data with I ≥ 2σ(I), S (all data) = 1.030, E_max, E_min = 0.48, −0.36 eÅ⁻³, respectively. CCDC 1848621.

Compound 3, crystal data; C₁₄H₂₁B₁₀PSe, M = 407.34, orthorhombic, Pbca, a = 9.9130(5), b = 18.9783(9), c = 20.7608(10) Å, U = 3905.8(3) ų, Z = 8, D_c = 1.385 Mg m⁻³, μ = 1.999 mm⁻¹, F(000) = 1632. 51443 data to θ_max = 31.49° were collected at 100.00(10) K on a Bruker X8 APEXII diffractometer using Mo-K_α X-radiation. A total of 6491 unique reflections (R_int = 0.0693) were used to solve and refine the structure, as for compound 1. R₁ = 0.0389, wR₂ = 0.0899 for data with I ≥ 2σ(I), S (all data) = 1.027, E_max, E_min = 0.81, −0.86 eÅ⁻³, respectively. CCDC 1848622.

Compound 4, crystal data; C₁₈H₁₀F₅PSe, M = 431.19, monoclinic, P2₁/c, a = 16.3024(12), b = 7.2293(5), c = 14.4191(11) Å, β = 104.365(4)°, U = 1646.2(2) ų, Z = 4, D_c = 1.740 Mg m⁻³, μ = 2.428 mm⁻¹, F(000) = 848. 489984 data to θ_max = 34.20° were collected at 100.00(10) K on a Bruker X8 APEXII diffractometer using Mo-K_α X-radiation. A total of 6774 unique reflections (R_int = 0.0396) were used to solve and refine the structure, as for compound 1. R₁ = 0.0262, wR₂ = 0.0611 for data with I ≥ 2σ(I), S (all data) = 1.056, E_max, E_min = 0.50, −0.46 eÅ⁻³, respectively. CCDC 1848623.

For 2 and 3, the cage C atoms were distinguished from B atoms by application of the Vertex-Centroid Distance (VCD) and Boron-Hydrogen Distance (BHD) methods [36,37,38].

4. Conclusions

The first example of a carborane with a catecholborolyl substituent, 1, was prepared and fully characterized, and was shown to enhance the catalysis of a Michael addition reaction by forming an FLP with PPh₃. A variety of carboranylphosphines were tested as FLP components in combination with B(C₆F₅)₃ as catalysts of a hydrosilylation reaction, with the strongly-basic V performing less well than the relatively weakly basic IVa, whilst the very weakly basic IX was completely ineffective. These results demonstrate that the ability to tune the Lewis acid and/or Lewis base strength of FLP components is critical in optimizing their use as catalysts and suggest that the electronic tuneability of carborane supports offers great potential in this respect.