Hydrogenation Versus Hydrosilylation: The Substantial Impact of a Palladium Capsule on the Catalytic Outcome

Abstract

A palladium capsule, made of three cavitands, namely P,P-dichlorido{5,17-bis[5-(diphenylphosphanyl)-4(24),6(10),12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4]arenyl-17-oxymthyl]-4(24),6(10),12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4]arene}palladium(II) (1), was synthetized by coordination of the corresponding diphosphinated ligand and the palladium precursor [PdCl₂(PhCN)₂] in 27% yield. The obtained P,P-chelate complex was fully characterized by elemental analysis, NMR and mass spectrometry. Molecular dynamics simulations carried out on the metallo-capsule showed the structure made by the three cavitands was slightly distorted over the 1 μs of the simulation. The evaluation of the palladium capsule 1 in the reaction between arylacetylenes and Et₃SiH in undried conditions unequivocally demonstrates a drastic change in chemoselectivity, with the formation of the partially hydrogenation product rather than the hydrosilylation products observed with complexes whose active center is more accessible, for instance [PdCl₂(PPh₃)₂].

1. Introduction

The chemistry of molecular containers is currently undergoing considerable development [1,2,3,4,5,6,7]. These objects are ideal for forming new host–guest complexes and can also be used to confine catalytic reactions, allowing precise control of catalytic reactions. Consisting of one or more concave pockets [8], for example from resorcin[4]arenes [9,10,11], these molecules share similarities with enzymes [12,13,14]. In this context, molecular capsules that can host a metal center are the preferred choice. Their cavity mimics the internal pocket of an enzyme, allowing reactive species to be stabilized [15] and substrates to be accommodated, thereby modifying the chemo-, regio- and stereoselectivity of the catalytic reaction [16,17,18].

Today, there are two distinct approaches in this field. The first approach focuses on catalysts trapped in supramolecular assemblies [19,20,21,22,23,24,25,26]. Examples include dimeric capsules based on glycoluril A [27] or resorcin[4]arene B [28] (Figure 1), which form during the catalytic process. The main advantage of this approach is its simplicity. These capsules can be easily obtained from simple precursors. Capsular edifices resulting from self-assembly are the seat of weak interactions that facilitate the reversible opening of the capsules formed, thus promoting the migration of molecules from inside the capsule to the outside (and vice versa). While self-assembled entities are most often labile, making catalyst recovery difficult, recent publications have convincingly demonstrated that spatial constraints within a self-assembled cavity can result in highly selective products, including stereoselective ones [29,30,31,32,33,34]. The supramolecular approach is the best way to discover and evaluate new models of “encompassing” catalysts, without the need to synthesize covalently constructed capsule units. The logical alternative approach is to use covalent capsules, such as capsule C described by Cram [35]. These require more effort to synthesize, but have the advantage of greater stability and easier recovery. Examples of covalent capsules are rarer [36], especially those built on the resorcin[4]arene platform [37].

Our approach was based on the use of a diphosphine comprising two resorcin[4]arene cavities covalently linked by a spacer. We synthesized it with the expectation that after chelation of a transition metal by the two phosphorus atoms, the two hemispherical resorcinarenyl units could be brought together to create a capsular complex. The platinum capsular complex D (Figure 2) comprised, at the time of its publication, the first covalent metallo-capsules based on resorcin[4]arene moieties to function as a catalyst [37].

We tested the metallo-capsule D in the hydroformylation of styrene. The active species was generated in situ with the help of SnCl₂ as a co-catalyst. Under optimized conditions, 24 h at 100 °C under 45 bar of CO/H₂, the platinum capsule catalyzed aldehyde formation with 55% conversion and 64% regioselectivity towards the branched product. By comparison, when the reaction was carried out using the trans-[PtCl₂(PPh₃)₂] complex, a conversion of 40% and a linear/branched ratio of 53:47 were observed. The higher activity observed with the capsular complex was undoubtedly due to the formation of intermediates whose metal center, located inside the capsule, adopted a distorted trigonal bipyramidal structure. This distortion facilitates hydride transfer to the coordinated olefin (proximity effect). The distorted trigonal bipyramidal structure is a direct result of the large P-Pt-P angle imposed by the ligand. Furthermore, steric interactions between the inner walls of the cavity and the catalytic center are slightly greater in the case of the linear Pt-alkyl intermediate. This is probably the reason for the regioselectivity observed in favor of the branched aldehyde [37].

We herein describe the synthesis and a catalytic application of the palladium capsule 1 in which the spacer between the two phosphinated resorcin[4]arene units is a cavitand (Figure 2).

2. Results and Discussion

2.1. Synthesis of the Palladium Capsule 1

The palladium capsule 1 was generated in two steps from the 5-(diphenylphosphanoyl)-17-hydroxy-4(24),6(10),12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4]arene (2) [37] and the 5,17-bis(bromomethyl)-4(24),6(10), 12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4]arene (3) [38] (Scheme 1). The alkylation of 2 with 3 in the presence of K₂CO₃ in DMF led to the formation of the tris-macrocyclic compound 4 in 54% yield after 86 h at 40 °C. The bisphosphine oxide 4 was reduced to the corresponding bisphosphine 5 with phenylsilane under reflux for 47 h. After treatment with methanol, the pure bisphosphane 5 was isolated in 93% yield. In its ³¹P NMR spectrum, the PPh₂ groups appeared as a singlet at −16.3 ppm.

Having in hand the bisphosphine 5, the formation of the P,P-chelate complex 1 was achieved under high-dilution conditions. Stoichiometric solutions of bisphosphine 5 and [PdCl₂(MeCN)₂] in CH₂Cl₂ were simultaneously added to a three-neck round-bottom flask containing pure CH₂Cl₂ over a period of three hours. In such high-dilution conditions, the palladium capsule 1 was isolated with a 27% yield. Its ³¹P NMR spectrum displayed a unique singlet at 11.6 ppm, which was consistent with a linear P-Pd-P arrangement. This was deduced from the variation in the ³¹P NMR signals of the phosphorus atoms before (in 5) and after (in 1) coordination (Δδ = 27.9 ppm) [39]. The ¹H and ¹³C NMR spectra were consistent with a Cs-symmetrical structure. This undoubtedly implies rapid oscillations on the NMR time scale (vide infra). The metallo-capsule has 210 hydrogen atoms, yet its ¹H NMR spectrum is relatively simple (Figure 3). The symmetry of the molecule allows us to observe only three AB patterns and three triplets, with each signal having intensities of eight and four protons, respectively, for the twelve OCH₂O and CH(C₆H₁₁) moieties, respectively. Similarly, there are only two and six visible singlets for the six and twelve aromatic protons of the upper and lower rims, respectively, of the three cavitands.

Furthermore, the formation of the P,P-chelate complex 1 is unambiguously demonstrated with its ESI-TOF MS spectrum, which displays a strong peak at m/z = 3055.36 and 3014.33, corresponding to [M − Cl + CH₃CN]⁺ and [M − Cl]⁺ mono-cations, respectively, and at m/z = 1526.65 and 1518.67, corresponding to [M − Cl + K]²⁺ and [M − Cl + Na]²⁺ bis-cations, respectively, with the expected and isotopic profiles.

2.2. Molecular Dynamics Study

To gain further insight into the conformational stability of the metallo-capsule 1, molecular dynamics simulations were performed. For this purpose, the palladium complex 1 was solvated in a 59.6 × 59.6 × 59.6 Å box containing 1500 molecules of dichloromethane. The system was then simulated over a period of 1 μs to observe its evolution at a temperature of 300 K. During this time, rapid motions of the twelve pentyl chains and slight distortions of the structure formed by the three resorcinarenyl units were observed (Figure 4).

Furthermore, the molecular dynamic simulations demonstrated the rapid diffusion of dichloromethane molecules from the external environment into the cavity formed by the three cavitands through two portals up to 10 Å in diameter. The average number of dichloromethane molecules hosted in this cavity was 3.15 (Figure 5).

2.3. Catalytic Application

The palladium-catalyzed reaction between an arylacetylene and a trialkylsilane leads generally to the hydrosilylated product in the form of a mixture of three isomers (β-E, β-Z and α) [40,41]. In the present work, the reaction was performed using a catalytic loading of 3 mol % in the presence of NaPF₆ as a co-catalyst under reflux of undried CH₂Cl₂ for 24 h (Scheme 2).

The reaction between phenylacetylene and triethylsilane carried out with the palladium capsule 1 showed that 49% of the acetylenic substrate was consumed. The detailed ¹H NMR analysis of the formed products clearly showed that hydrosilylated products 7 were formed in low proportions, with styrene being the main product (58% of the formed products). The β/α regioselectivity of silylated products was 1.3/1 (

Table 1. Palladium-catalyzed hydrosilylation of phenylacetylene [a].

Entry	Palladium Complex	Acetylenic (6)	Conversion (%) [b]	Product Distribution [b]
1	1		49	18	0	14	58
2	1		11	10	0	5	85
3	1		26	32	0	6	62
4	9		100	69	13	15	3
5	10		70	57	4	11	28
6 [c]	10 + D2O		47	48	3	9	39 (D:H ratio = 83:17)
7 [d]	10 + molecular sieve		75	59	23	8	10

[a] Palladium complex (1.5 μmol, 3.0 mol%), NaPF₆ (0.30 mg, 1.8 μmol, 3.6 mol%), arylacetylene (5.5 μL, 50.0 μmol), Et₃SiH (20 μL, 125.0 μmol), CH₂Cl₂ (1.0 mL), reflux, 24 h. [b] Determinate by ¹H NMR [42,43]. [c] Addition of D₂O (500 μmol, 9 μL). [d] Reaction carried out in the presence of molecular sieve 4 Å.

, entry 1). When the runs were repeated with 4-fluorophenylacetylene and 4-methoxyphenylacetylene, it was also possible to form, with lower conversions, mainly the partially hydrogenated products in proportions of 85 and 62%, respectively (Table 1, entries 2 and 3).

The use of [PdCl₂(PPh₃)₂] (9; Figure 6) as a catalyst resulted in a faster reaction, a full conversion of phenylacetylene and a notable, quasi-quantitative shift in chemoselectivity towards the formation of hydrosilylated compounds with a β/α ratio of 5.5 (Table 1, entry 4). Repeating the run with the palladium complex 10 (Figure 6) bearing two sterically hindered resorcinarenyl phosphines resulted in an intermediate conversion of 70% with lower chemoselectivity toward the hydrosilylated compounds (β/α ratio of 5.5) and formation of the partially hydrogenated product (28% of the formed products) (Table 1, entry 5). Adding a drop of D₂O to the catalytic mixture reduced the conversion to 47% and drastically modified the product’s outcome. In these conditions, the partially hydrogenated compound was formed with a chemoselectivity of 39%, and the β/α ratio (5.5) of the silylated product was not affected (Table 1, entry 6). As expected, the addition of a molecular sieve reduced the amount of formed styrene to 10% and slightly increased the conversion of the catalytic reaction to 75% (Table 1, entry 7).

A catalytic test with complex 10 in the absence of NaPF₆ resulted in a lower conversion of phenylacetylene (28%; see Table S1, entry 5 in Supplementary Materials) than in the presence of the sodium salt (conversion of 70%; Table 1, entry 5). This result highlights the beneficial role of NaPF₆ on the reactivity of complex 10, showing that the mechanism of hydrosilylation of arylacetylenes is more complex than the one generally proposed involving the formation of [L₂Pd⁽⁰⁾] as active species (active species I) [41], which is generated by reduction of [L₂Pd^(II)Cl₂] with Et₃SiH [44] (Scheme 3).

In our case, the reaction of [L₂Pd^(II)Cl₂] with NaPF₆ generated the [L₂Pd^(II)Cl]⁺ cation (II), which in the presence of Et₃SiH led to the formation of the palladium hydride [L₂Pd^(II)H]⁺ (III), reported as an active species in the hydrosilylation of arylacetylenes [45]. In the presence of trace H₂O, the unsaturated cationic species III can react to form intermediate IV, which after reaction with Et₃SiH leads to the active species for the partially hydrogenation reaction, the [L₂Pd^(II)(H)₂] (V) derivative [43]. Furthermore, the experiment proved that, in the absence of Et₃SiH, acetophenone was formed as a unique product as the result of water addition on phenylacetylene (13% of the phenylacetylene was converted into acetophenone using palladium complex 10; see Table S1, entry 4 in Supplementary Materials). This demonstrates that the formation of active species V, responsible for the partially hydrogenation product, requires the presence of Et₃SiH.

The catalytic results clearly indicate that in the case of resorcinarenyl ligands, complexes 1 and 10, a higher proportion of partially hydrogenated product 8 was formed. This is probably due to the fact that the lifetime of intermediate III, the active species for hydrosilylation, is long enough to react with water and form [L₂Pd^(II)(H)₂] (V). The latter may either evolve, in the case of complex 9 bearing two PPh₃ moieties, into [L₂Pd⁽⁰⁾] (I) and catalyze the hydrosilylation reaction, or react quickly, in the case of macrocyclic ligands (complexes 1 and 10), with arylacetylene previously hosted in the cavity of a resorcin[4]arene unit (spatial proximity) to lead to the partial hydrogenation products. With the PPh₃ ligand (complex 9), this supramolecular assistance does not exist, which makes the reaction with arylacetylene slower and therefore leaves the intermediate V enough time to eliminate H₂ and form the active species for the hydrosilylation reaction (I).

Moreover, the sterically constrained environment clearly influences the regioselectivity of the hydrosilylated products. Catalytic results show that, when 1 is employed, the β/α ratio of the silylated compounds is lower and there is relatively more formation of α product. Given its more compact nature, α product is formed more readily when the catalytic reaction is conducted within the palladium capsule 1. Such a proportion of the Markovnikov product (α 7) is unusual for a palladium catalyst and is generally observed when ruthenium or cobalt complexes are employed [46].

3. Materials and Methods

General Procedure: All manipulations involving phosphorus derivatives were performed in Schlenk-type flasks under dry argon. Solvents were dried by conventional methods and distilled immediately before use. CDCl₃ was passed down a 5 cm-thick alumina column and stored under nitrogen over molecular sieves (4 Å). Routine ¹H, ¹³C{¹H} and ³¹P{¹H} spectra were recorded with an AC 500 MHz Bruker FT spectrometer. The ¹H and ¹³C NMR spectra were referenced to residual protonated solvents (δ = 7.26 and 77.16 ppm, respectively). The ³¹P NMR spectroscopic data are given relative to external H₃PO₄. Chemical shifts and coupling constants are labeled in ppm and Hz, respectively. Elemental analyses were performed by the Service de Microanalyse, Institut de Chimie, Université de Strasbourg. Mass spectra were recorded on a Bruker MicroTOF spectrometer (ESI-TOF). 5-(Diphenylphosphanoyl)-17-hydroxy-4(24),6(10),12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4]arene (2) [37], 5,17-bis(bromomethyl)-4(24),6(10),12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4]arene (3) [38] and trans-P,P-dichlorido-bis[5-diphenylphosphanyl-4(24),6(10),12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4]arene]palladium(II) (10) [47] were prepared by literature procedures.

3.1. Synthesis of 5,17-bis[5-(Diphenylphosphanoyl)-4(24),6(10),12(16),18(22)-tetramethylene-dioxy-2,8,14,20-tetrapentylresorcin[4]arenyl-17-oxymethyl]-4(24),6(10),12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4]arene (4)

Cavitand 2 (290 mg, 0.28 mmol) and K₂CO₃ (155 mg, 1.12 mmol) were suspended in DMF (15 mL). After 0.5 h, the resorcinarene 3 (138 mg, 0.14 mmol) was added and the resulting mixture was heated at 40 °C for 86 h. The reaction mixture was then allowed to cool down to room temperature and water (40 mL) was added. The aqueous layer was extracted with CH₂Cl₂ (3 × 40 mL). The combined organic layers were washed with water (5 × 50 mL) and brine (50 mL), then dried over MgSO₄. The solvent was then evaporated under reduced pressure and the crude product purified by column chromatography (CH₂Cl₂/Et₂O, 9:1 v/v; Rf = 0.57: CH₂Cl₂/Et₂O, 9:1 v/v), yield 212 mg (54%). ¹H NMR (500 MHz, CDCl₃): δ = 7.70–7.86 (m, 8H, arom. CH, P(O)Ph₂), 7.42–7.35 (m, 12H, arom. CH, P(O)Ph₂), 7.38 (s, 2H, arom. CH, resorcinarene), 7.20 (s, 2H, arom. CH, resorcinarene), 7.15 (s, 2H, arom. CH, resorcinarene), 7.07 (s, 4H, arom. CH, resorcinarene), 6.87 (s, 2H, arom. CH, resorcinarene), 6.48 (s, 2H, arom. CH, resorcinarene), 6.38 (s, 4H, arom. CH, resorcinarene), 5.93 and 4.51 (AB spin system, 8H, OCH₂O, ²J_HH = 7.5 Hz), 5.84 and 4.50 (AB spin system, 8H, OCH₂O, ²J_HH = 7.5 Hz), 4.80 (t, 4H, CHCH₂, ³J_HH = 8.0 Hz), 4.78 (s, 4H, OCH₂C), 4.71 (t, 4H, CHCH₂, ³J_HH = 8.0 Hz), 4.64 (t, 4H, CHCH₂, ³J_HH = 8.0 Hz), 4.59 and 4.22 (AB spin system, 8H, OCH₂O, ²J_HH = 7.5 Hz), 2.33–2.21 (m, 16H, CHCH₂CH₂), 2.17–2.10 (m, 8H, CHCH₂CH₂), 1.46–1.32 (m, 72H, CH₂CH₂CH₂CH₃), 0.94 (t, 12H, CH₂CH₃, ³J_HH = 7.0 Hz), 0.93 (t, 12H, CH₂CH₃, ³J_HH = 7.0 Hz), 0.90 (t, 12H, CH₂CH₃, ³J_HH = 7.0 Hz); ¹³C{¹H} NMR (126 MHz, CDCl₃): δ = 156.52–114.43 (arom. C’s), 100.29 (s, OCH₂O), 100.12 (s, OCH₂O), 100.02 (s, OCH₂O), 65.94 (s, OCH₂C), 36.83 (s, CHCH₂), 36.81 (s, CHCH₂), 36.53 (s, CHCH₂), 32.25 (s, CH₂CH₂CH₃), 32.19 (s, CH₂CH₂CH₃), 32.17 (s, CH₂CH₂CH₃), 30.25 (s, CHCH₂), 30.21 (s, CHCH₂), 30.11 (s, CHCH₂), 27.78 (s, CHCH₂CH₂), 27.71 (s, CHCH₂CH₂), 22.85 (s, CH₂CH₃), 22.82 (s, CH₂CH₃), 14.28 (s, CH₂CH₃), 14.25 (s, CH₂CH₃); ³¹P{¹H} NMR (202 MHz, CDCl₃): δ = 23.1 (s, P(O)Ph₂) ppm. Elemental analysis calculated (%) for C₁₈₂H₂₁₀O₂₈P₂ (2907.54): C 75.18, H 7.28; found: C 74.92, H 7.50.

3.2. Synthesis of 5,17-bis[5-(Diphenylphosphanyl)-4(24),6(10),12(16),18(22)-tetramethylene-dioxy-2,8,14,20-tetrapentylresorcin[4]arenyl-17-oxymethyl]-4(24),6(10),12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4]arene (5)

Cavitand 4 (177 mg, 0.06 mmol) was diluted in phenylsilane (3 mL) and the resulting suspension was refluxed for 47 h. The volatiles were then evaporated under reduced pressure and the crude product was dissolved in CH₂Cl₂ (1 mL) and precipitated with methanol (5 mL). The white precipitate was filtered, washed with methanol (3 × 3 mL) and dried in a vacuum to produce 5 as a white solid, yield 163 mg (93%). ¹H NMR (500 MHz, CDCl₃): δ = 7.32–7.24 (m, 20H, arom. CH, PPh₂), 7.20 (s, 2H, arom. CH, resorcinarene), 7.19 (s, 2H, arom. CH, resorcinarene), 7.15 (s, 2H, arom. CH, resorcinarene), 7.08 (s, 4H, arom. CH, resorcinarene), 6.91 (s, 2H, arom. CH, resorcinarene), 6.51 (s, 2H, arom. CH, resorcinarene), 6.36 (s, 4H, arom. CH, resorcinarene), 5.94 and 4.49 (AB spin system, 8H, OCH₂O, ²J_HH = 7.5 Hz), 5.80 and 4.36 (AB spin system, 8H, OCH₂O, ²J_HH = 7.0 Hz), 4.82 (t, 4H, CHCH₂, ³J_HH = 8.0 Hz), 4.77 (s, 4H, OCH₂C), 4.75 and 4.08 (AB spin system, 8H, OCH₂O, ²J_HH = 7.5 Hz), 4.72 (t, 4H, CHCH₂, ³J_HH = 8.5 Hz), 4.70 (t, 4H, CHCH₂, ³J_HH = 8.3 Hz), 2.31–2.22 (m, 16H, CHCH₂CH₂), 2.19–2.12 (m, 8H, CHCH₂CH₂), 1.48–1.32 (m, 72H, CH₂CH₂CH₂CH₃), 0.94 (t, 12H, CH₂CH₃, ³J_HH = 7.5 Hz), 0.91 (t, 24H, CH₂CH₃, ³J_HH = 7.0 Hz); ¹³C{¹H} NMR (126 MHz, CDCl₃): δ = 157.21–114.96 (arom. C’s), 100.07 (s, OCH₂O), 99.97 (s, OCH₂O), 99.42 (s, OCH₂O), 66.07 (s, OCH₂C), 36.90 (s, CHCH₂), 36.84 (s, CHCH₂), 36.77 (s, CHCH₂), 32.24 (s, CH₂CH₂CH₃), 32.22 (s, CH₂CH₂CH₃), 32.13 (s, CH₂CH₂CH₃), 30.15 (s, CHCH₂), 27.78 (s, CHCH₂CH₂), 27.77 (s, CHCH₂CH₂), 27.69 (s, CHCH₂CH₂), 22.88 (s, CH₂CH₃), 22.87 (s, CH₂CH₃), 22.85 (s, CH₂CH₃), 14.30 (s, CH₂CH₃), 14.27 (s, CH₂CH₃), 14.25 (s, CH₂CH₃); ³¹P{¹H} NMR (202 MHz, CDCl₃): δ = −16.3 (s, PPh₂) ppm. Elemental analysis calculated (%) for C₁₈₂H₂₁₀O₂₆P₂ (2875.54): C 76.02, H 7.36; found: C 75.74, H 7.24.

3.3. Synthesis of P,P-Dichlorido{5,17-bis[5-(diphenylphosphanyl)-4(24),6(10),12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4]arenyl-17-oxymethyl]-4(24),6(10),12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4]arene}palladium(II) (1)

A solution of bisphosphine 5 (212 mg, 0.07 mmol) in CH₂Cl₂ (100 mL) and a solution of [PdCl₂(MeCN)₂] (28 mg, 0.07 mmol) in CH₂Cl₂ (100 mL) were added simultaneously over a period of 3 h in CH₂Cl₂ (1300 mL) and the resulting solution was stirred for an additional period of 16 h at room temperature. The solvent was then evaporated under reduced pressure and the crude product was purified by column chromatography (CH₂Cl₂; Rf = 0.87: CH₂Cl₂). The resulting solid was dissolved in CH₂Cl₂ (0.5 mL) and precipitated with Et₂O (25 mL), and the yellow solid was collected and dried in a vacuum, yield 61 mg (27%). ¹H NMR (500 MHz, CDCl₃): δ = 7.70–7.66 (m, 8H, arom. CH, PPh₂), 7.38–7.34 (m, 6H, arom. CH, PPh₂), 7.33 (s, 2H, arom. CH, resorcinarene), 7.29–7.25 (m, 6H, arom. CH, PPh₂), 7.19 (s, 2H, arom. CH, resorcinarene), 7.18 (s, 2H, arom. CH, resorcinarene), 7.11 (s, 4H, arom. CH, resorcinarene), 6.95 (s, 2H, arom. CH, resorcinarene), 6.52 (s, 2H, arom. CH, resorcinarene), 6.29 (s, 4H, arom. CH, resorcinarene), 5.95 and 4.51 (AB spin system, 8H, OCH₂O, ²J_HH = 7.5 Hz), 5.80 and 4.49 (AB spin system, 8H, OCH₂O, ²J_HH = 7.0 Hz), 5.27 and 4.13 (AB spin system, 8H, OCH₂O, ²J_HH = 8.0 Hz), 4.87 (s, 4H, OCH₂C), 4.80 (t, 4H, CHCH₂, ³J_HH = 8.0 Hz), 4.77 (t, 4H, CHCH₂, ³J_HH = 8.0 Hz), 4.60 (t, 4H, CHCH₂, ³J_HH = 8.0 Hz), 2.27-2.16 (m, 24H, CHCH₂CH₂), 1.47–1.31 (m, 72H, CH₂CH₂CH₂CH₃), 0.93 (t, 12H, CH₂CH₃, ³J_HH = 7.0 Hz), 0.93 (t, 24H, CH₂CH₃, ³J_HH = 7.5 Hz); ¹³C{¹H} NMR (126 MHz, CDCl₃): δ = 157.99–115.28 (arom. C’s), 100.63 (s, OCH₂O), 100.54 (s, OCH₂O), 99.72 (s, OCH₂O), 58.88 (s, OCH₂C), 36.92 (s, CHCH₂), 36.86 (s, CHCH₂), 36.65 (s, CHCH₂), 32.20 (s, CH₂CH₂CH₃), 32.19 (s, CH₂CH₂CH₃), 31.98 (s, CH₂CH₂CH₃), (s, CHCH₂), 30.48 (s, CHCH₂), 30.22 (s, CHCH₂), 30.14 (s, CHCH₂), 27.74 (s, CHCH₂CH₂), 27.72 (s, CHCH₂CH₂), 27.56 (s, CHCH₂CH₂), 22.88 (s, CH₂CH₃), 22.85 (s, CH₂CH₃), 14.28 (s, CH₂CH₃), 14.27 (s, CH₂CH₃); ³¹P{¹H} NMR (202 MHz, CDCl₃): δ = 11.6 (s, PPh₂) ppm. MS (ESI-TOF): m/z = 3055.36 [M − Cl + CH₃CN]⁺, 3014.33 [M − Cl]⁺, 1526.65 [M − Cl + K]²⁺, 1518.67 [M − Cl + Na]²⁺ (expected isotopic profiles). Elemental analysis calculated (%) for C₁₈₂H₂₁₀O₂₆P₂PdCl₂•Et₂O (3126.99): C 71.44, H 7.09; found: C 71.52, H 7.15.

3.4. Molecular Dynamics Simulations

The different systems were simulated by classic molecular dynamics (MD) using AMBER.22 GPU software [48] in which the potential energy U is empirically described by a sum of bond, angle and dihedral deformation energies and pairwise additive 1-6-12 (electrostatic + van der Waals) interactions between non-bonded atoms.

$$\begin{array}{c}U={\displaystyle \sum _{bonds}{k}_b}{(r-r_0)}^2+{\displaystyle \sum _{angles}{k}_{\theta }}{(\theta -{\theta }_0)}^2+{\displaystyle \sum _{dihedrals}{V}_n}\left[1+cos(n\varphi -\gamma )\right]\\ +{\displaystyle \sum _{i=1}^{N-1}{\displaystyle \sum _{j=i+1}^N\left[\frac{A_{ij}}{R_{ij}^{12}}-\frac{B_{ij}}{R_{ij}^6}-\frac{q_i{q}_j}{{\epsilon }_0{r}_{ij}}\right]}}\end{array}$$

The initial structure of the capsule was constructed by “hand” and optimized by semi-empirical method. The molecule was then solvated by 1500 molecules of CH₂Cl₂ (box size is around 59.6 × 59.6 × 59.6 Å).

Force field parameters come from the GAFF force field [49] and atomic charges were obtained using the RESP methodology (B3LYP geometry optimization and HF single-point methods for RESP calculation and 6-31G(d,p) basis set for H,C,P,Cl and SDD for Pd) [50]. Palladium parameters were from Kurt [51]. For CH₂Cl₂ we used the model of Kollmann [52]. Cross terms in van der Waals interactions were constructed using the Lorentz-Berthelot rules. 1–4 van der Waals and 1–4 electrostatic interactions were scaled by a factor of 1.2 and 2, respectively. The MD simulations were performed at 300 K, starting with random velocities. All simulations were carried out using 3D Periodic boundary conditions. An atom-based cut-off of 12 Å for non-bonded interactions was applied and long-range electrostatics were calculated using the Ewald summation method in the particle mesh Ewald (PME) approximation, while a long-range correction for vdW interaction was applied [53]. The SHAKE algorithm from AMBER was used to fix any bonds involving hydrogen atoms.

The different systems were first equilibrated by 0.5 ns of dynamics in the NPT ensemble, followed by 0.5 ns of dynamics in the NVT ensemble. Finally, production runs of 1 μs in the NVT ensemble were simulated. The system temperature was controlled using a Berendsen thermostat with a time step of 1 ps. Pressure was controlled using a Berendsen barostat. A time step of 2 fs was used to integrate the equations of motion via the Verlet leapfrog algorithm. Trajectories were saved every 20 ps. Snapshots along the trajectory were taken using VMD 1.9.4 software [54]. The trajectories were analyzed using our MDS software [55]. The average number of solvent molecules in the cavity was obtained by radial density functions (RDFs) calculated between the center of mass of the capsule and the carbon atoms of the dichloromethane solvent molecules. Integration of the RDF curves (along the dynamics) at 5.5 Å (approximate radius of the cavity) gave an average number of 3.15 ± 0.4.

3.5. Catalytic Tests

Under argon, the palladium complex (1.5 μmol, 3.0 mol%) was dissolved in CH₂Cl₂ (undried, 1.0 mL). NaPF₆ (0.30 mg, 1.8 μmol, 3.6 mol%, weighted in air), phenylacetylene (5.5 μL, 50.0 μmol) and Et₃SiH (20 μL, 125.0 μmol) were added and the reaction mixture was refluxed for 24 h. The solution was then concentrated in a vacuum and analyzed by ¹H NMR. The spectra were compared with the literature [42,43].

4. Conclusions

Here we describe the two-step synthesis of 5,17-bis[5-(diphenylphosphanyl)-4(24),6(10),12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4]arenyl-17-oxymthyl]-4(24),6(10),12(16),18(22)-tetramethylenedioxy-2,8,14,20-tetrapentylresorcin[4] arene (5) starting from two synthons previously developed in our laboratory. The strategy of coordinating the precursor [PdCl₂(PhCN)₂] with the two phosphorus atoms of 5 will produce a backbone replenishment of the three cavitands, resulting in the metallo-capsule 1. Under ultra-high dilution, the P,P-chelate 1 was isolated in 27% yield. In agreement with ¹H and ¹³C NMR analysis, which revealed Cs pseudo-symmetry, molecular dynamics simulations realized on the palladium capsule 1 showed that the skeleton of the molecule oscillated rapidly without altering its structure.

The palladium capsule proved to be a surprisingly effective catalyst in the reaction between arylacetylenes and Et₃SiH under undried conditions, leading mainly to partially hydrogenated derivatives rather than the expected hydrosilylated products. This is in contrast to the use of [PdCl₂(PPh₃)₂], which typically results in the formation of the hydrosilylated products. Moreover, the sterically constrained environment clearly influenced the regioselectivity of the hydrosilylated products. Catalytic results show that, when 1 is employed, the β/α ratio of the silylated compounds is lower and there is relatively more formation of α product. Given its more compact nature, α product is formed more readily when the catalytic reaction is conducted within the palladium capsule 1. The β/α ratio of the silylated products is indisputably lower when the P,P-chelate 1 is employed than when analogs devoid of capsular units are used. These observations are the result of two factors: (i) the longer lifetime of the intermediate resulting from the first step of the catalytic cycle, the [L₂PdH]⁺ intermediate, and its lower reactivity with arylacetylenes, which allows, in the presence of water, the formation of the [L₂Pd(H)₂] active species of the partially hydrogenated products and (ii) steric interactions between the capsule wall and the intermediately formed Pd-alkyl units result in the formation of a higher proportion of the more compact α-hydrosilylated product.

Further studies will exploit the potential of metallo-capsules based on phosphinated cavitands in the chemoselective formation of carbon-heteroatom bond-forming reactions.