Control of charge transfer in donor/acceptor metal-organic frameworks

Acc Chem Res. 2013 Feb 19;46(2):248-57. doi: 10.1021/ar300102t. Epub 2012 Nov 5.


Charge transfer (CT) of D(0)A(0) ↔ D(δ+)A(δ-) not only involves an electron transfer from D to A, but also generates a new spin set of S = 1/2 spins with an exchange interaction. Therefore, the control of CT in multidimensional frameworks could be an efficient way to design electronically/magnetically functional materials. The use of redox-active metal complexes as D and/or A building blocks expands the variety of such D/A frameworks with the formulation of D(m)A(n) (m, n ≥ 1), permitting the design of donor/acceptor metal-organic frameworks (D/A-MOFs). This Account summarizes our ongoing research on the design of D/A-MOFs and on the systematic control of CT in such D/A-MOFs toward the discovery of unique electronic/magnetic materials exhibiting nontrivial phenomena. For this purpose, the D/A combinations of carboxylate-bridged paddlewheel-type diruthenium(II,II) complexes ([Ru(2)(II,II)]) that act as one-electron (1e(-)) donors and polycyanoorganic acceptors such as 7,7,8,8-tetracyano-p-quinodimethane (TCNQ) and N,N'-dicyanoquinodiimine (DCNQI) have been chosen. Even in the covalently bonded motif, the CT in this system is systematically dependent on the intrinsic ionization potential (I(D)) and electron affinity (E(A)) of the D and A units, respectively, which is controllable by chemical modification of the D/A units. As we consider the energy difference between the HOMO of D and the LUMO of A (ΔE(H-L)(DA)) instead of hν(CT) ∝ |I(D) - E(A)|, the neutral (N) and ionic (I) states can be defined as follows: (i) the D/A materials with ΔE(H-L)(DA) > 0 (i.e., the LUMO level of A is higher than the HOMO level of D) should be neutral, and (ii) complexes adopted when ΔE(H-L)(DA) < 0 are, meanwhile, ionic. Materials located near ΔE(H-L)(DA) ≈ 0, that is, at the boundary between the N and I phases, are candidates for the N-I transition driven by external stimuli such as temperature, pressure, and photoirradiation. Even in the ionic state, two distinct states could be isolated for the D(2)A type: (ii-1) the 1e(-) transferred D(2)A-MOFs provide mixed-valence systems of D(+)D(0)A(-) possibly involving intervalence CT, which produce magnetic correlations via radical A(-) units, and (ii-2) when the 2e(-) reduced form of A (e.g., TCNQ(2-)) is energetically favored beyond the on-site Coulomb repulsion on A, the oxidation state of D(+)(2)A(2-) is produced, for which magnetic measurements reveal a paramagnetic state attributed to the isolated D(+) units. The interspatial Coulombic interaction is another factor in determining the charge distribution in materials, which is related to the spatial Coulombic stability of D/A packing and possibly yields a mixture of N and I domains when it is more advantageous to get Coulombic gain than in the uniform N or I phase. Such a phase could be observed at the boundary between N and I phases involving the N-I transition. These charge-distributed states/phases are systematically demonstrated in a D/A-MOF system made by the combination of [Ru(2)(II,II)] and TCNQ/DCNQI; however, we immediately recognize the charge distribution of D/A-MOF only by understanding the nature of the starting D/A units. The present D/A-MOF system should be an intriguing platform to look for new functionalities with synergistic correlations among charge, spin, and lattice.