Pathway selection in the self-assembly of Rh4L4 coordination squares under kinetic control

Abstract

Pathway selection principles in reversible reaction networks such as molecular self-assembly have not been established yet, because achieving kinetic control in reversible reaction networks is more complicated than in irreversible ones. In this study, we discovered that coordination squares consisting of cis-protected dinuclear rhodium(II) corner complexes and linear ditopic ligands are assembled under kinetic control, perfectly preventing the corresponding triangles, by modulating their energy landscapes with a weak monotopic carboxylate ligand (2,6-dichlorobenzoate: dcb^-) as the leaving ligand. Experimental and numerical approaches revealed the self-assembly pathway where the cyclization step to form the triangular complex is blocked by dcb^-. It was also found that one of the molecular squares assembled into a dimeric structure owing to the solvophobic effect, which was characterized by nuclear magnetic resonance spectroscopy and single-crystal X-ray analysis.

Fig. 1. Self-assembly of the Rh₄L₄ squares in this study.

Schematic representation of kinetically controlled self-assembly of the Rh₄L₄ squares from cis-protected dinuclear Rh(II) complex (Rh²⁺: [Rh₂(DAniF)₂]²⁺) and linear dicarboxylate ligand (L^2–: 1^2– or 2^2–) in solution. When the metal source is [Rh(CH₃CN)₄]²⁺, where CH₃CN is the leaving ligand, Rh₃L₃ triangle and Rh₄L₄ square complexes were produced in a 1:2 ratio. In contrast, when the self-assembly was carried out using Rh(dcb)₂ as the metal source (dcb^–: 2,6-dichlorobenzoate), dcb^– modulates the energy landscape of the self-assembly so that the formation of the Rh₃L₃ triangle is prevented, producing the Rh₄L₄ square only. The axial ligands on the dinuclear Rh(II) centers are omitted for clarity.

Fig. 2. ¹H NMR spectra of the self-assembly of the Rh₄1₄ square.

¹H NMR spectra (500 MHz, CDCl₃, 298 K, aromatic region) of the self-assembly of the Rh₄L₄ squares under various conditions ([1^2–] = [Rh²⁺] = 1 mM). a Self-assembly from [Rh(CH₃CN)₄](BF₄)₂ and 1^2– in CDCl₃ at 298 K, giving the Rh₃1₃ triangle and the Rh₄1₄ square in 23% and 39% yields, respectively. Asterisks indicate the carbon satellite of CHCl₃. b Addition of n-Bu₄N·dcb in a mixture of the Rh₃1₃ triangle and the Rh₄1₄ square obtained from the self-assembly of [Rh(CH₃CN)₄](BF₄)₂ and 1^2– in CDCl₃ after convergence. The Rh₃1₃ triangle was converted into the Rh₄1₄ square at 298 K assisted by dcb^–, although heating at 100 °C for 2 days is necessary without dcb^– (Supplementary Fig. 14). c Self-assembly of the Rh₄1₄ square from Rh(dcb)₂ and 1^2– in CDCl₃ at 298 K to produce the Rh₄1₄ square in a 65% yield without formation of the Rh₃1₃ triangle during self-assembly. The yields were determined based on the internal standard.

Fig. 3. Self-assembly mechanism of the Rh₄2₄ square under kinetic control.

a Three types of possible chain intermediates in the self-assembly of M₄L₄ square. Type I: M_mL_m+1, Type II: M_m+1L_m+1(dcb), Type III: M_m+2L_m+1(dcb)₂. b Plots of the existence ratios of the substrates and products in the self-assembly of the Rh₄2₄ square from Rh(dcb)₂ and 2^2– in CDCl₃ at 298 K. [Rh]₀ = [2^2–]₀ = 0.86 mM. c Plots of the (〈n〉, 〈k〉) values in the n-k map of the Rh₄2₄ square (red filled circles). Green crosshairs indicate the (n, k) values of the chain intermediates. The three types of chain intermediates, Types I, II, and III, are plotted on each straight line. The brown number indicates m in each type of oligomer in a. The definitions of 〈n〉 and 〈k〉 values are shown in the main text (Eqs. 1 and 2). The data in (b) and (c) are shown as the average of the three runs of QASAP with standard errors. d Comparison of the existence ratio of the substrates and the products between QASAP (red) and NASAP (blue). e Comparison of the n-k plot between QASAP (red) and NASAP (blue). Both (d) and (e) indicate that the numerical simulation results reproduce the experimental counterparts well. f Dominant self-assembly pathways of the Rh₄2₄ square. Red arrows indicate the major self-assembly pathway. (a,b,c) indicates Rh_a2_b(dcb)_c. The numbers above the reaction arrows indicate the elementary reactions with high net reaction frequency listed in Supplementary Table 4. Longer Type II oligomers than (4,4,1), such as (5,5,1) and (6,6,1), are produced during the self-assembly, although they are not involved in the major self-assembly pathway (Supplementary Fig. 13). g Two possible pathways to produce triangle (3,3,0) from Type II intermediate (3,3,1) (left) and from Type II oligomers with more than three Rh²⁺ units (4+x, 4+x, 1) (x ≥ 0) (right). The green sphere indicates the leaving ligand(s) (dcb^– or two molecules of CH₃CN). h A plausible key structure in the triangular formation through associative ligand exchange process. Steric repulsion caused by ArCO₂^– (in green) in the cyclic intermediate and transition state of the triangular formation process would prevent the cyclization.

Fig. 4. Dimerization of the Rh₄1₄ square in solution.

a Supramolecular dimer formation of the Rh₄1₄ square by the solvophobic effect. b Crystal structure of [Rh₄1₄(dmso-S)₄]₂. Two Rh₄1₄ squares engaged each other are shown in red and blue. DMSO molecules axially coordinating to the Rh(II) centers are colored in green. c ¹H NMR spectra (500 MHz, 298 K, aromatic region) of (Rh₄1₄)₂ in CD₃NO₂/CDCl₃ (9:1 (v/v)) and Rh₄1₄ in CDCl₃. d ¹H DOSY spectrum of (Rh₄1₄)₂ in CD₃NO₂/CDCl₃ (9:1 (v/v)).