Pushing the Thermodynamic and Kinetic Limits of Near-Infrared Emissive CrIII Complexes in Photocatalysis

Abstract

Photoactive Cr^III complexes are typically based on polypyridine coordination environments, exhibit red luminescence, and are good photo-oxidants but have modest photoreducing properties. We report new Cr^III complexes with anionic chelate ligands that enable color-tunable near-infrared luminescence and red-light-driven photoreduction reactions involving elementary steps that are endergonic up to 0.5 eV. Improving the metal-ligand bond covalency rather than more established approaches such as optimizing ligand field strength and coordination geometry is the underlying molecular design concept to achieve this favorable behavior. Our analysis suggests an intricate interplay between productive but slow endergonic photoinduced electron transfer and energy-wasting charge recombination rooted in cage escape effects, which could be generally important for photocatalysis. Our work also suggests the occurrence of doublet-doublet annihilation, a process that seems to have been largely neglected in current research on photoactive Cr^III complexes but which could provide a mechanistic entry point into the widely used process of photochemical upconversion, typically based on triplet-triplet annihilation. Overall, this work conceptually advances the current state of the art of photoactive Cr^III complexes in terms of molecular design, luminescence, and photoredox behavior. More generally, it informs photochemistry in terms of elucidating the limits of light-to-chemical energy conversion efficiency and the value of long-lived excited states in complexes of earth-abundant transition metals.

1

Molecular structures and key electrochemical and photophysical properties of relevant published and new Cr^III complexes. The electron-deficient and strongly red luminescent (a) [Cr­(dqp)₂]³⁺ and (b) [Cr­(ddpd)₂]³⁺, for which ligand field strength and optimized coordination geometry were key design criteria. The new electron-rich near-infrared luminescent [Cr­(^RBTP)₂]⁻ complexes (c,d), with greater metal–ligand bond covalency as a key molecular design concept (R = CF₃ or tBuPh). Nonluminescent [Cr­(^MePDP)₂]⁻ complex (e). E _red is the electrochemical potential for one-electron reduction in V vs SCE, λ_em is the emission band maximum in solution at room temperature.

2

(a) Synthesis of the bis­(1H-1,2,4-triazol-5-yl)­pyridine (^RBTP) ligands, where R is either a (tert-butyl)­phenyl (tBuPh) or a trifluoromethyl (CF₃) group. X = Cl or OH in the case of R = tBuPh, and X = OH in the case of R = CF₃. (b) Synthesis of the [Cr­(^RBTP)₂]⁻ complexes.

3

Solid-state X-ray crystal structure of (a) [Cr­(^CF3BTP)₂]⁻ and (b) [Cr­( ^tBuPhBTP)₂]⁻ viewed from two different angles. Disordered atoms, hydrogen atoms, and solvent molecules have been omitted for clarity.

4

(a) UV–vis absorption (solid traces) and emission (dashed traces) spectra of (nBu₄N)­[Cr­(^CF3BTP)₂] (blue, λ_ex = 350 nm) and (nBu₄N)­[Cr­( ^tBuPhBTP)₂] (red, λ_ex = 300 nm). (b) Magnification of the absorption spectra between 400 and 800 nm. (c) Transient UV–vis absorption spectra at different delay times for [Cr­( ^tBuPhBTP)₂]⁻ (40 μM, laser pump pulses at 355 nm, 30 mJ/pulse, and integration time of 1.5 μs). Inset: decay of the excited-state absorption (ESA) at 680 nm and recovery of the ground state bleach (GSB) at 385 nm. (d) Transient UV–vis absorption spectra at different delay times for [Cr­(^CF3BTP)₂]⁻ (40 μM, laser pump pulses at 355 nm, 30 mJ/pulse, integration time of 1.5 μs). Inset: decay of the ESA at 468 nm and recovery of the GSB at 345 nm. All measurements were performed in deoxygenated acetonitrile at room temperature.

5

(a) Simplified Tanabe–Sugano diagram for an octahedral d³ configuration, with the ²E/²T₁ spin-flip excited states (green) and the ligand field-dependent ⁴T₂ excited state (brown) included, while all other states are not shown. The dashed vertical line represents the case of [Cr­(dqp)₂]³⁺, for which 10 Dq/B = 38. (b) Resulting shift of the potential energy surfaces of the states involved upon decreasing the 10 Dq/B ratio in the presence of π-donor anionic (deprotonated) triazole ligand units. The potential energy surfaces and some relevant microstates are marked in black (⁴A₂, ground state), green (²E, spin-flip state), and brown (⁴T₂, metal-centered state).

6

(a) Cyclic voltammograms of [Cr­( ^tBuPhBTP)₂]⁻ (red) and [Cr­(^CF3BTP)₂]⁻ (blue), recorded from 0.5 mM solutions in Ar-flushed acetonitrile in the presence of 0.1 M nBu₄NPF₆ as a supporting electrolyte. Scan rate: 0.2 V/s. Working electrode: glassy carbon. Counter electrode: silver wire. Reference electrode: KCl saturated calomel (SCE). (b,c) Latimer diagrams including the ground states and the photoactive ²E/²T₁ excited states of [Cr­( ^tBuPhBTP)₂]⁻ and [Cr­(^CF3BTP)₂]⁻.

7

Rate constants for bimolecular electron transfer (k _q) from selected electron donors (EDs, table on the right) to photoexcited [Cr­( ^tBuPhBTP)₂]⁻ (black squares) as a function of the reaction free energy (ΔG _ET). The red line describes the trend obtained by fitting the data with the empirical Rehm–Weller equation (Supporting Information). The best fit was obtained with a diffusion constant k _d of 3.0 × 10¹⁰ M^–1 s^–1, a parameter m accounting for the dissociation of the Cr^III–ED encounter complex of 0.025, and a free activation energy ΔG _ET ^‡ (0) of 0.23 eV in acetonitrile at the point where ΔG _ET = 0.

8

(a) Energy level diagram for the exergonic (ΔG _ET = −0.9 eV) photoinduced electron transfer (ET) from tetrakis­(dimethylamino)­ethylene (TDAE, ED₁) to the excited state of the photocatalyst (*­[Cr]⁻) and following in-cage charge recombination (iCR) in the encounter complex; (b) energy level diagram for the endergonic (ΔG _ET = +0.5 eV) photoinduced electron transfer from N,N-dimethylaniline (DMA, ED₉). The free energy of the charge recombination is still negative (−2.0 eV), but with a higher absolute value than the previous case (−0.6 eV); (c) classical Marcus parabola describing the driving-force dependence of the rate constant for in-cage charge recombination (k _iCR). The unwanted back electron transfer from the reduced photocatalyst [Cr]^2– to ED_9^•+ occurs in the inverted regime further away from the maximum of the parabola, and therefore can be slower than the respective back electron transfer from [Cr]^2– to ED_1^•+ . For simplicity, short symbols representing the reduced complex are shown, but the reduction is ligand-based, as displayed in the Latimer diagram in Figure b,c.

9

(a) Photocatalytic reductive debromination of α-bromoacetophenone upon irradiation with either a 405 or 632 nm LED and (b) oxidation of α-terpinene by singlet oxygen following irradiation at 632 nm in the presence of [Cr­( ^tBuPhBTP)₂]⁻. The product yields (and the substrate conversions reported in parentheses) were determined by HPLC measurements (using suitable calibration curves) in the case of (a) or by ¹H NMR spectroscopy referred to an internal standard for (b).

10

(a) Dependence of the lifetime of the ²E/²T₁ excited state of [Cr­( ^tBuPhBTP)₂]⁻ (τ) in Ar-flushed acetonitrile at 25 °C on the excitation pulse energy (P_exc) in mJ, displayed here in the form of its reciprocal square root. The duration of the excitation pulses was approximately 10 ns. The red circles were obtained from a 7.0 mM solution that was excited at 532 nm, whereas the blue circles were recorded from a 0.1 mM solution that was excited at 355 nm. Both solutions had equal absorbance (0.4) at the respective excitation wavelength. The ²E/²T₁ lifetime was detected by monitoring the disappearance of the excited-state absorption at 680 nm. The solid black lines are linear regression fits to the experimental data. (b,c) Possible bimolecular processes affecting the ²E/²T₁ lifetime: (b) doublet–doublet annihilation (DDA) and (c) excited-state disproportionation (ESD) followed by charge recombination (CR) to the quartet ground state. For simplicity and for consistency with Figure , short symbols representing the reduced and oxidized complex are shown in (c), as ESD could in principle involve both metal-centered and ligand-based redox events; see also Figure b,c.