Threefold symmetric trimetallic macrocycles formed via the Weak-Link Approach

Abstract

The synthesis and characterization of a new threefold symmetric hemilabile phosphino-alkylthioether ligand are described. This ligand can be used in combination with Rh(I) and Ir(I) precursor complexes to prepare 40-membered macrocyles with threefold symmetry via the Weak-Link Approach. Synthesis and characterization of two such structures are reported along with a single crystal x-ray diffraction analysis of one of the key intermediates in the case of Rh(I). This is a demonstration of the viability of the Weak-Link Approach for preparing structures other than bimetallic macrocycles and suggests that it could be generalized for a wide range of higher symmetry structures through appropriate hemilabile ligand design.

Recent developments in the field of supramolecular chemistry have resulted in the establishment of several new synthetic strategies for forming multimetallic macrocyclic architectures (1–6). Our group has introduced and focused on developing one of these strategies, which is referred to as the Weak-Link Approach (Scheme S1). This strategy allows one to prepare bimetallic macrocycles from simple transition metal and ligand precursors (7–12), and it is based on reactions between flexible symmetrical hemilable ligands (13–16) and transition metal precursors. Typically, macrocyclic intermediates that contain both strong (S-M) and weak (W-M) ligand–metal bonds (Scheme S1) are initially formed in very high and usually quantitative yield. For the systems studied thus far, ligands that consist of flexible phosphinoalkyl ether (7), thioether (12), or amino (11) bidentate ligand functionalities connected with rigid aromatic spacers (R) have been used. Once condensed, macrocyclic intermediates are formed, the weak ligand–metal bonds in these intermediates can be cleaved by means of simple ligand substitution reactions to afford open macrocyclic ring structures. Alternatively, both condensed and opened macrocycles can be used to construct more complex three-dimensional architectures such as cylinders (8) and trilayer metallocyclophanes (9).

Scheme 1.

Although the generality of the weak-link methodology has been demonstrated with respect to twofold symmetric hemilabile ligands and transition metals, the application of this synthetic approach to the preparation of macrocyclic structures with more than two metals has not been demonstrated. Herein, we report the first application of the Weak-Link Approach to the formation of trimetallic (17–20) rhodium and iridium macrocycles from a flexible threefold symmetric hemilabile ligand.

Methods and Materials

General Procedures.

All reactions were performed under a nitrogen atmosphere in reagent grade solvents by using standard Schlenk or dry-box techniques (21). All other solvents were purified by published methods (22, 23). Chemicals were purchased from Aldrich, unless otherwise mentioned, or prepared by literature methods, as referenced below. ¹H and ³¹P{¹H} NMR spectra were recorded on a Varian Mercury 300 MHz spectrometer by using deuterated solvents and H₃PO₄ as internal and external references, respectively. Solution infrared spectra were recorded on a Nicolet-520 spectrometer by using NaCl cells with 0.1-mm spacers. Elemental analyses were obtained from Quantitative Technologies (Whitehouse, NJ).

Synthesis of TPB(SCH₂CH₂PPh₂)₃ (TPB, 1,3,5-triphenylbenzene) (2).

Solid NaSMe (0.60 g, 8.63 mmol) was added in one portion to a solution of 1,3,5-Tris(4-bromophenyl)benzene (1) (1.53 g, 2.82 mmol) in dimethylformamide (50 ml). The reaction mixture was stirred for 1 h at 150°C. Additional NaSMe (0.60 g, 8.63 mmol) was added, heating was continued for 1 h, and then a third portion of NaSMe (0.60 g, 8.63 mmol) was added to the reaction mixture. The mixture was heated at 150°C for another 6 h, cooled to 40°C, and diluted with acetic acid (2 ml) and water (60 ml). The product forms as a pale-gray precipitate, which was filtered, washed with water, and dried in vacuum to afford analytically pure 1,3,5-Tris(4-mercaptophenyl)benzene (1.01 g, 89%). ¹H NMR: (CD₂Cl₂) δ 3.52 (s, 3 H, SH), 7.38 (m, 6 H, Ph-H), 7.56 (m, 6 H, Ph-H) AA′BB′, 7.67 (s, 3 H, Ph-H). C₂₄H₁₈S₃: calcd 71.60% C, 4.51% H, 23.89% S; found 71.53% C, 4.52% H, 23.93% S.

Neat CH₂=CHPPh₂ (0.9 ml, 3.62 mmol) was added to a solution of 1,3,5-Tris(4-mercaptophenyl)benzene (0.67 g, 2.35 mmol) in tetrahydrofuran (50 ml). A catalytic amount of azobisisobutyronitrile (AIBN) (5.0 mg, 0.03 mmol) was added to the reaction mixture, which was subsequently heated to reflux and stirred. After 2 h, more AIBN (5.0 mg, 0.03 mmol) was added to the mixture, which was stirred for an additional 6 h at reflux. The reaction mixture was cooled, and all volatiles were removed under vacuum to afford a white foamy solid, which was washed with EtOH (3 × 20 ml) and dried under vacuum to give analytically pure 2 (0.678 g, 90%). ¹H NMR: (CD₂Cl₂) δ 2.44 (m, 6 H, PCH₂), 3.00 (m, 6 H, SCH₂), 7.23–7.45 (complex m, 36 H, Ph-H), 7.63 (d, 6 H, Ph-H, J = 8.4 Hz), 7.77 (s, 3 H, Ph-H). ³¹P{¹H} NMR: (CD₂Cl₂) δ −16 (s). C₆₆H₅₇P₃S₃: calcd 76.27% C, 5.53% H, 8.94% P, 9.26% S; found 75.49% C, 5.40% H, 9.07% P, 9.14% S.

Synthesis of [{TPB(SCH₂CH₂PPh₂)₃}₂M₃](BF₄)₃ (4a,b; M = Rh, Ir).

A solution of Rh(acac)(NBD) (NBD, norbornadiene) (acac, acetylacetonate) (100 mg, 0.34 mmol) in CH₂Cl₂ (5 ml) was added to a solution of [Ph₃C]BF₄ (132 mg, 0.34 mmol) in CH₂Cl₂ (5 ml) to produce a clear yellow mixture. Then a CH₂Cl₂ solution of 2 (117 mg, 0.11 mmol, 5 ml) was added dropwise to the mixture over 5 min. After the reaction mixture was stirred for 20 min at ambient temperature, the resulting orange–yellow solution contained primarily [{TPB(SCH₂CH₂PPh₂)₃}Rh₃(NBD)₃](BF₄)₃ (3a), as indicated by a single resonance [δ 57 (d, J_Rh-P = 154 Hz)] in the ³¹P{¹H} NMR analysis of the mixture. Then a solution of 2 (second equiv; 117 mg, 0.11 mmol) in CH₂Cl₂ (5 ml) was added dropwise over 5 min. The mixture was stirred for an additional 20 min to give a clear yellow solution, which was layered with Et₂O (50 ml) to precipitate 4a (305 mg, 94%) as fine yellow needles. Compound 4a: ¹H NMR: (CD₂Cl₂) δ 2.60–2.82 (br s, SCH₂CH₂PPh₂, 24H), 7.20–7.55 (complex m, P(C₆H₅)₂, 84H), 7.70 (s, Ph-H, 6H). ³¹P{¹H} NMR: (CD₂Cl₂) δ 66 (d, J_Rh-P = 165 Hz). Electrospray–MS: [[M] − 3BF]³⁺ calcd = 795.8, expt = 795.4 m/z. C₁₃₂H₁₁₄B₃F₁₂P₆Rh₃S₆: calcd 59.88% C, 4.34% H, 7.02% P; found 60.08% C, 3.95% H, 6.88% P. Crystals of 4a suitable for x-ray diffraction analysis were obtained by slow diethyl ether vapor diffusion into a saturated solution of 4a in CH₂Cl₂/C₆H₆/CH₃CN (10:5:1) at room temperature.

A mixture of [Ir(1,5-COD)Cl]₂ (1,5-COD, 1,5-cyclooctadiene) (50 mg, 0.074 mmol) and AgBF₄ (29 mg, 0.15 mmol) in CH₂Cl₂ (5 ml) was stirred for 30 min to give a deep red solution and gray precipitate of AgCl. The solution was filtered through a short Celite plug. Then a solution of 2 (52 mg, 0.050 mmol) in CH₂Cl₂ (5 ml) was added dropwise over 5 min to the filtrate. After the mixture was stirred for 20 min at ambient temperature, the resulting orange solution contained primarily [{TPB(SCH₂CH₂PPh₂)₃}Ir₃(1,5-COD)₃](BF₄)₃ (3b), as indicated by a single resonance [δ 48 (s)] in the ³¹P{¹H} NMR analysis of the mixture. Complex 4b (145 mg, 89%) was synthesized from 3b in a similar fashion as 4a. Compound 4b: ¹H NMR: (CD₂Cl₂) δ 2.53 (br s, CH₂PPh₂, 12H), 2.85 (br s, SCH₂, 12H), 7.10–7.70 (complex m, P(C₆H₅)₂, 90H). ³¹P{¹H} NMR: (CD₂Cl₂) δ 49 (s). Electrospray–MS: [[M] − 3BF]³⁺ calcd = 885.0, expt = 885.0 m/z.

Synthesis of {TPB(SCH₂CH₂PPh₂)₃}₂M₃(CO)₃Cl₃ (5a,b; M = Rh, Ir).

A slow stream of CO was bubbled through a solution of 4a (10 mg, 0.004 mmol) and [Me₄N]Cl (5 mg) in CD₂Cl₂ (1.5 ml) for 10 s. The color of the solution immediately changed from orange–yellow to pale yellow. The resulting solution was filtered from the white precipitate of [Me₄N]BF₄, and the solvent was removed under gentle vacuum (≈10 mm Hg) to give 5a as a pale-yellow solid (9 mg, 95%). Compound 5a: ¹H NMR: (CD₂Cl₂) δ 3.05 (br s, CH₂PPh₂, 12H), 3.50 (br s, SCH₂, 12H), 7.20–7.80 (complex m, P(C₆H₅)₂, 90H). ³¹P{¹H} NMR: (CD₂Cl₂) δ 24 (d, J_Rh-P = 124 Hz). Fourier-transformed (FT) IR: (CH₂Cl₂): ν_CO = 1,974 cm⁻¹(vs). The susceptibility of 5a to losing CO and Cl⁻ precluded its characterization by mass spectrometry. Complex 5b was synthesized from 4b in a similar fashion as 5a. Compound 5b: ¹H NMR: δ (CD₂Cl₂) 3.11 (br s, CH₂PPh₂, 12H), 3.47 (br s, SCH₂, 12H), 7.05–7.89 [complex m, P(C₆H₅)₂, 90H]. ³¹P{¹H} NMR: (CD₂Cl₂) δ 20 (br s). FTIR: (CH₂Cl₂): ν_CO = 1,961 cm⁻¹(vs).

General Procedure for X-Ray Structure Determination of 4a.

Data were collected on a Bruker (Madison, WI) SMART APEX charge-coupled device diffractometer. The structure was solved by using direct methods and standard difference map techniques. Absorption correction was made by sadabs (Version 2.01, Bruker). In addition to the cation and three BF anions in the crystal structure of 4a, there are three CH₂Cl₂ and C₆H₆ disordered solvent molecules. A treatment by squeeze (24) was used for solvate molecules. Correction of the x-ray data for 4a by squeeze (339 electron/cell) was very close to the required value (336 electron/cell) for the above-mentioned solvate molecules. Three Ph rings in the cation and some F atoms in the anions are disordered on two positions with occupancy in the ranges 0.23–0.65 and 0.41–0.55, respectively. All non-H atoms were refined with anisotropic displacement coefficients, except for the disordered atoms. All hydrogen atoms were placed in the structure factor calculation at idealized positions. All software and sources of the scattering factors are contained in the shelxtl (5.10) program package.

Crystal data: [120.0(0.1) K] C₁₄₁H₁₂₆B₃Cl₆F₁₂P₆Rh₃S₆, FW = 2,980.46, triclinic, P-1, a = 17.3238(11) Å, b = 19.2861(12) Å, c = 21.2459(13) Å, α = 82.9140(10)°, β = 70.6660(10)°, γ = 79.9810(10)°, Z = 2, V = 6579.4(7) ų, d(calcd) = 1.504 g cm⁻¹, μ = 0.730 mm⁻¹, 29,650 unique reflections, R1 = 0.0880, wR2 = 0.2071.

Results and Discussion

Synthesis of TPB(SCH₂CH₂PPh₂)₃ (2).

Ligand 2 was prepared by a two-step procedure starting from 1,3,5-tris(4-bromophenyl)benzene (25) (1) in 72% yield (Scheme S2). Hydrothiolation of 1 with an excess of NaSMe in dimethylformamide followed by acid quenching afforded 1,3,5-Tris(4-mercaptophenyl)benzene, which was reacted with vinyldiphenylphosphine in the presence of azobisisobutyronitrile to yield 2. Spectroscopic data for 2 are completely consistent with its proposed structure. Ligand 2 is conformationally flexible (C_3ν symmetry in solution), with the three thioetherphosphine functionalities in fixed diverging spatial orientations.

Scheme 2.

Synthesis and Characterization of the Condensed Macrocycles [{TPB(SCH₂CH₂PPh₂)₃}₂M₃](BF₄)₃ (4a,b; M = Rh, Ir).

Recently, we reported that the thioether-based ligand 1,4-(PPh₂CH₂CH₂S)₂C₆H₄ cleanly reacts with a Rh(I) precursor generated via the reaction between [Rh(COE)₂Cl]_x (COE, cyclooctene) and AgBF₄ to form the bimetallic condensed macrocyclic complex [{1,4-(Ph₂PCH₂CH₂S)₂C₆H₄}₂Rh₂]⁺² in >90% yield (Eq. 1) (12).

In contrast, when ligand 2 is reacted with the product from the reaction between [Rh(COE)₂Cl]_x AgBF₄ in CH₂Cl₂ under similar reaction conditions, a complex mixture of products containing both rhodium metal and ligand 2 is obtained, as evidenced by ¹H and ³¹P{¹H} NMR spectroscopy. Although we were unable to separate and/or isolate individual products from that reaction, some of the resonances and coupling patterns in the ¹H and ³¹P{¹H} NMR spectra are very similar to those of target complex 4a (see text below, Scheme S3), which suggests that complex 4a is formed in this reaction along with oligomeric side products. Attempts to optimize the reaction conditions, to increase the yield of 4a, by using different solvents (tetrahydrofuran, acetone, MeOH), high-dilution (0.02 mN concentration of reactants), and low-temperature (−78°C) techniques were unsuccessful. Therefore, we sought to develop a synthetic strategy, which allows one to construct three-dimensional condensed macrocycles of type 4 in a more controlled manner.

Scheme 3.

Rhodium-containing condensed macrocycle 4a was synthesized in two steps, beginning with the reaction between ligand 2 (0.33 equiv) and the product of the 1:1 reaction between Rh(acac)(NBD) and [Ph₃C]BF₄ at room temperature in CH₂Cl₂. This results in the formation of trinuclear complex [{TPB(SCH₂CH₂PPh₂)₃}Rh₃(NBD)₃]⁺³ (3a), which has been spectroscopically characterized. It is readily identified by its characteristic ³¹P{¹H} NMR spectrum, which shows a single doublet at δ 57 (J_Rh-P = 154 Hz) (26). Complex 4a was synthesized in nearly quantitative yield by slow addition of ligand 2 (0.33 equiv) to the solution containing 3a in CH₂Cl₂. In contrast, when Ir(acac)(1,5-COD) was used as an iridium source for the generation of complex 3b, the reaction led to a complex mixture of products with 3b as a major component, as evidenced by ³¹P{¹H} NMR spectroscopy [δ 48 (s)]. However, 3b was synthesized in nearly quantitative yield by first reacting [Ir(1,5-COD)Cl]₂ with AgBF₄ in CH₂Cl₂ in 1:1 fashion, followed by the addition of ligand 2 (0.33 equiv) at room temperature. Intermediate 3b also was observed by ³¹P{¹H} NMR spectroscopy and then converted without further purification to 4b by slow addition of ligand 2 (0.33 equiv) to a solution of 3b in CH₂Cl₂.

Compounds 4a and b were isolated by precipitation through the addition of diethyl ether to the reaction medium. They are indefinitely stable at room temperature in both the solid and solution states. Both complexes are soluble in polar organic solvents (e.g., MeCN and CH₂Cl₂). Their ¹H NMR spectra are consistent with highly symmetric compounds (C₃ symmetry) and exhibit characteristic downfield shifts (≈0.2 ppm, compared with free ligand) for the inner ethylene bridges, which are due to the coordination of the metal to both the phosphorus and sulfur atoms of ligand 2. Consistent with their proposed structures, the ³¹P{¹H} NMR spectra of 4a,b each exhibits a single sharp resonance (a doublet at δ 64 with J_Rh-P = 162 Hz for 4a and a singlet at δ 49 for 4b). The coupling constant in the case of 4a is highly diagnostic of cis-phosphine cis-thioether coordination geometry around Rh(I) (27). Significantly, the absence of oligomeric species at the end of the reactions involving 3a and b (Scheme S3) strongly suggests that the ability of ligand 2 to preorganize the metal centers through chelation into rigid templates (3a,b) is a key element of the process. This trigonal template can then react with a second equivalent of the ligand in a highly cooperative manner (because of symmetry and spatial complementarity) to form the desired complexes 4a,b.

The solid-state structure of 4a (Fig. 1) was determined by an x-ray crystallographic study and is consistent with the proposed solution structure. The three rhodium atoms are bridged by the thioetherphosphine groups of the TPB(SCH₂CH₂PPh₂)₃ ligand 1, which binds in an η²-fashion to each of the rhodium atoms. Each rhodium center exhibits slightly distorted square planar geometry with two cis-phosphine and two cis-thioether moieties; the angles around Rh range from 85.26° to 98.38°. The Rh(I) centers are arranged in the form of a triangle to form an internal cavity that resembles a trigonal prism. The metal–metal distances range from 12.242 Å to 15.671 Å. No BF anions or solvent molecules were located within the cavity between the triphenylbenzene fragments. Note that the terminal phenyl rings and the middle ring of the triphenylbenzene group are twisted with respect to each other, which is evident by the significant torsion angles of the adjacent phenyl rings (25.1–38.0°). The average interplanar distance between the parallel planar phenyl rings of the triphenylbenzene fragments is 3.62 Å, which is similar to that observed in related metallocyclic complexes (28).

Figure 1.

Stick representations of [{TPB(SCH₂CH₂PPh₂)₃}₂Rh₃]⁺³ from the crystal structure of 4a. Hydrogens, counter-ions (BF), and solvent molecules are omitted for clarity. Rhodium, purple; sulfur, orange; phosphorus, green; carbon, gray.

Synthesis and Characterization of the Opened Macrocycles {TPB(SCH₂CH₂PPh₂)₃}₂M₃(CO)₃Cl₃ (5a,b; M = Rh, Ir).

Condensed macrocyclic complexes 4a,b undergo clean expansion by selectively cleaving their M–S bonds with [Me₄N]Cl under CO (1 atm), which results in the quantitative formation of the neutral macrocycles 5a,b (Scheme S4). Complexes 5a,b have been characterized by ¹H and ³¹P{¹H} NMR spectroscopy, which are consistent with their high symmetry. For example, the ³¹P{¹H} NMR spectra of each complex exhibit a single resonance [δ 24 (d, J_Rh-P = 124 Hz) for 5a (Fig. 2) and δ 20 (s) for 5b] for the six magnetically equivalent phosphorous atoms of trans PPh₂ groups of the ligand. In addition, the FT IR spectroscopy of 5a and b exhibit characteristic ν_CO bands for their terminal CO ligands at 1,973 and 1,961 cm⁻¹, respectively. The positions of these bands correlate well with mononuclear Rh and Ir complexes with near-identical coordination environments (29). Complexes 5a,b are stable in CH₂Cl₂ solution under CO (1 atm) at room temperature for several days. However, the exposure of complex 5a to high vacuum or prolonged storage in the solid state under nitrogen resulted in its conversion to the Cl⁻ salt, analogous to 4a.

Scheme 4.

Figure 2.

³¹P{¹H} NMR spectra of 4a and 5a in CH₂Cl₂. No other complexes are observed in the range from δ − 80 to δ + 80.

Conclusion

In conclusion, we have demonstrated that the Weak-Link Approach can be successfully applied to the synthesis of three-dimensional 40-membered inorganic macrocycles using a novel hemilabile ligand with threefold symmetry. This work demonstrates the viability of the weak-link approach for preparing structures other than bimetallic macrocycles and suggests that the approach could be generalized for a wide variety of higher symmetry structures through appropriate hemilabile ligand design. The ability to rationally tailor the sizes and shapes of such structures will directly impact their utility for host guest chemistry and shape selective catalysis.

Abbreviations

NBD

norbornadiene

1,5-COD

1,5-cyclooctadiene

TPB

1,3,5-triphenylbenzene