Exchange controlled triplet fusion in metal-organic frameworks

Abstract

Triplet-fusion-based photon upconversion holds promise for a wide range of applications, from photovoltaics to bioimaging. The efficiency of triplet fusion, however, is fundamentally limited in conventional molecular and polymeric systems by its spin dependence. Here, we show that the inherent tailorability of metal-organic frameworks (MOFs), combined with their highly porous but ordered structure, minimizes intertriplet exchange coupling and engineers effective spin mixing between singlet and quintet triplet-triplet pair states. We demonstrate singlet-quintet coupling in a pyrene-based MOF, NU-1000. An anomalous magnetic field effect is observed from NU-1000 corresponding to an induced resonance between singlet and quintet states that yields an increased fusion rate at room temperature under a relatively low applied magnetic field of 0.14 T. Our results suggest that MOFs offer particular promise for engineering the spin dynamics of multiexcitonic processes and improving their upconversion performance.

Fig. 1. Spin dynamics of the triplet fusion process and a strategy for enabling singlet–quintet (SQ) coupling.

a, Schematic illustration of TT pair dynamics. More than half of the TT pairs form quintet states initially, thus control of SQ coupling can substantially impact triplet fission and fusion efficiencies. b, Energy levels and spin states of nine eigenstates of TT pairs as a function of the distance between two triplet excitons. As shown in the inset, in a separated TT pair, the eigenstates are mixed SQ or triplet–quintet states governed by the zero-field splitting defined by the parameters D and E (ref. ). The intertriplet exchange interaction, J, dominates as the triplets get closer, eventually yielding pure-spin eigenstates. c, A resonant magnetic field can drive an oscillation between the singlet and quintet states. For SQ splitting on the order of D, the spin transition time is ∼ℏπ/D. d,e, Spatial description of triplet fusion in molecular solids (d) and MOFs (e). Moderate applied magnetic fields μB ≈ D reshuffle the singlet, triplet and quintet character of TT pairs in the shaded magnetic field effect (MFE) regions, generating both the conventional Merrifield MFE and potentially SQ resonances. TT states with weaker exchange splitting J ≈ D may be precursors to exciton formation in MOFs. In a typical densely packed molecular solid, however, triplets migrate rapidly relative to ℏπ/D and fusion typically occurs from a triplet pair with exchange splitting J ≫ D. This means that, in molecular solids, SQ resonances may only be observed under very high applied magnetic fields μ_ΒB ≫ D and low temperature, if at all.

Fig. 2. Magnetic field effect (MFE) on triplet–triplet pair eigenstate energies and triplet fusion rate for three different values of the intertriplet exchange coupling.

a–c, MFE on the eigenstates of TT pairs with different intertriplet exchange interactions: J = 0 μeV (a), J = –6.1 μeV (b) and J = –100 μeV (c). d–f, MFE on the normalized triplet fusion rates that are proportional to singlet characteristics of TT pairs for J = 0 μeV (d), J = –6.1 μeV (e) and J = –100 μeV (f). When there is no exchange interaction (a,d), the singlet-like TT pair population is modulated by the intratriplet dipole–dipole interaction and Zeeman interaction. These are Merrifield-type MFE characteristics, as the model assumes J = 0. When a strong exchange interaction is present (c,f), each eigenstate represents the total spin states of TT pairs. The fusion rate increases when quintet states mix with singlet states near the avoided level crossings. When the exchange coupling is comparable to intratriplet dipole coupling (b,e), each state is not a pure spin state, but contains a dominant spin character. The fusion rate increases near the avoided crossing of singlet-dominant and quintet-dominant states. The resonance positions deviate from their expected doubling of the field strength, as modulated by the relative sign of D and J. Note that these calculations assume a single crystal with zero-field splitting parameter D = 8.5 μeV, whereas the experimental MFE curve was obtained from polycrystalline materials. Note also that J is a function of TT separation, as described in Fig. 1.

Fig. 3. Portions of the structures of NU-1000 and NU-901, a triplet sensitizer, and their photophysical properties.

a–e, Structure of NU-1000 (a), NU-901 (b), oxo-Zr₆ SBU (c), H₄TBAPy (linker) (d) and PtOEP (triplet sensitizer) (e). Note the unique placements of the ligand molecules in the MOFs that are distinct from molecular solids. PtOEP molecules are incorporated into the large pores of the MOFs. f, Photoluminescence spectra of NU-1000, NU-901 and PtOEP. Blue and red curves represent NU-1000 fluorescence (Fl.) and phosphorescence (Ph.), respectively. Yellow and purple curves show NU-901 fluorescence and phosphorescence spectra, respectively. The PtOEP phosphorescence (green curve) is higher in energy than those of either MOF, ensuring efficient Dexter energy transfer from PtOEP to the MOFs. g, Schematic of the triplet-fusion-based photon upconversion process in the NU-1000:PtOEP system. h, The upconverted emission spectrum of NU-1000:PtOEP excited with a λ = 532 nm laser. i, Pump-power dependence of upconverted emission exhibiting a slope change from 2 to 1. Such a power dependence transition is observed when the dominant decay mechanism for triplet excitons shifts from a first-order process to TT fusion, and coincides with maximizing the upconversion efficiency. The upconversion threshold intensity is 35 mW cm⁻².

Fig. 4. Magnetic field effects on upconverted emission.

a, NU-1000:PtOEP has an unique MFE with two additional distinct peaks around 0.14 and 0.33 T. b,c, NU-901:PtOEP (b) and a physical mixture of H₄TBAPy:PtOEP (10 wt%) (c) show the conventional MFE for triplet fusion, confirming the critical role of structure in NU-1000 for enabling anomalous MFE behaviour with NU-1000:PtOEP. Error bars are the standard deviation of the mean and are averaged from five, seven and four independent sweeps for a, b and c, respectively.