Outpacing conventional nicotinamide hydrogenation catalysis by a strongly communicating heterodinuclear photocatalyst

Abstract

Unequivocal assignment of rate-limiting steps in supramolecular photocatalysts is of utmost importance to rationally optimize photocatalytic activity. By spectroscopic and catalytic analysis of a series of three structurally similar [(tbbpy)₂Ru-BL-Rh(Cp*)Cl]³⁺ photocatalysts just differing in the central part (alkynyl, triazole or phenazine) of the bridging ligand (BL) we are able to derive design strategies for improved photocatalytic activity of this class of compounds (tbbpy = 4,4´-tert-butyl-2,2´-bipyridine, Cp* = pentamethylcyclopentadienyl). Most importantly, not the rate of the transfer of the first electron towards the Rh^III center but rather the rate at which a two-fold reduced Rh^I species is generated can directly be correlated with the observed photocatalytic formation of NADH from NAD⁺. Interestingly, the complex which exhibits the fastest intramolecular electron transfer kinetics for the first electron is not the one that allows the fastest photocatalysis. With the photocatalytically most efficient alkynyl linked system, it is even possible to overcome the rate of thermal NADH formation by avoiding the rate-determining β-hydride elimination step. Moreover, for this photocatalyst loss of the alkynyl functionality under photocatalytic conditions is identified as an important deactivation pathway.

Fig. 1. Mechanisms and investigated compounds for nicotinamide reduction.

a Depiction of the photocatalytic and thermal catalytic hydrogenation of nicotinamides by the selected class of rhodium catalysts^–,. Photocatalysis: (I) light-driven two electron reduction of Rh^III, (II) oxidative addition of a proton, (III) hydride transfer; Thermal catalysis: (I_th) coordination of formate and elimination of CO₂, (III) hydride transfer. b Molecular structure of the investigated photocatalysts 2, 4 and 5.

Fig. 2. Synthesis route of complexes 1, 2, 3 and 4.

(i) (1) 5-ethynyl-1,10-phenanthroline, THF, diisopropylamine, [Pd(PPh₃)₄], CuI, 60 °C, 24 h; (2) KCN, dichloromethane/H₂O, rt, 1 h; (ii) (1) 5-azido-1,10-phenanthroline, CuSO₄, sodium ascorbate, H₂O/dichloromethane, rt, 24 h; (2) KCN, dichloromethane/H₂O, rt, 1 h; (iii) [Rh(Cp*)Cl₂]₂, dichloromethane, rt, 24 h.

Fig. 3. Redox properties of the complexes.

Cyclic voltammograms of separate 1 mM acetonitrile solutions of 1 (blue), 2 (orange), 3 (green) and 4 (red) at room temperature with nBu₄NPF₆ as supporting electrolyte (0.1 M). An Ag wire is used as quasi reference electrode, Pt wire as the counter electrode and glassy carbon as the working electrode. All data referenced against Fc⁺/Fc; scan rate = 100 mV s⁻¹.

Fig. 4. Optical properties of the complexes 1, 2, 3 and 4.

a UV/vis absorption (solid) and emission (dashed) spectra (λ_exc = 450 nm) of the samples. b, e Nanosecond transient absorption spectra of 1 (blue), 2 (orange), 3 (green) and 4 (red) collected upon excitation at 470 nm in acetonitrile. c, f Time-resolved emission spectra of 1 (blue), 2 (orange), 3 (green) and 4 (red) (experimental data) in acetonitrile excited at 470 nm. d, g Emission decay profiles of the transient species and normalized transient absorption kinetic traces recorded in the region of the ground-state bleach at 450 nm and the corresponding fit (cyan, dashed). The respectively fitted emission and transient absorption lifetimes are indicated in the plots. For the color codes, see c and f.

Fig. 5. Formate-driven nicotinamide reduction.

a TONs (solid dots, dotted lines to guide the eye) and TOFs (hollow dots, dotted lines to guide the eye) during formate-driven catalysis at 45 °C of complexes 2 (orange), 4 (red), and 5 (gray) (H₂O/acetonitrile (9/1, v/v), 50 mM NaHCO₂, 5 µM catalyst, 250 µM NAD⁺). b Average TOF (within 90 min or until substrate limitation) during formate-driven catalysis of complexes 2 (orange), 4 (red), and 5 (gray) at temperatures ranging from 25 to 50 °C (dotted lines to guide the eye; error bars represent standard deviation, n = 2 independent measurements).

Fig. 6. Photochemical reduction of the heterodinuclear complexes.

Absorbance at 680 nm during irradiation (470 nm, 54 mW/cm²) of deaerated photocatalyst solutions (0.12 M TEA, acetonitrile/H₂O (1/9, v/v)). Complexes 2, 4 and 5 are represented by orange, red and gray solid lines, respectively.

Fig. 7. Photocatalysis by the heterodinuclear complexes.

Temperature dependent photocatalysis of complexes 2 (orange), 4 (red) and 5 (gray). TON as solid dots, TOF as hollow dots (dotted lines guide the eye; error bars represent standard deviation, n = 2 individual measurements for all panels). a Temperature dependent TOF within the first 10 min of photocatalysis and average TOF of 2, 4 and 5 (within 90 min or until substrate limitation) in formate-driven catalysis (blue) for comparison. b TON after 90 min of photocatalysis for the three complexes (acetonitrile/H₂O (1/2, v/v), 0.12 M TEA, 0.1 M NaH₂PO₄, 5 µM catalyst, 250 µM NAD⁺). c Photocatalysis at 25 °C. d Photocatalysis at 45 °C.

Fig. 8. Resonance Raman spectroscopy of complexes 1–4 and references.

RR spectra of 1 and 2 in acetonitrile upon excitation at 405 nm (a) and 473 nm (d). RR spectra of 3 and 4 in acetonitrile excited at 405 nm (b) and 473 nm (e). RR spectrum of [(tbbpy)₂Ru(phen)]²⁺, excited at 405 nm (c) and 473 nm (f) for comparison. Characteristic Raman modes assigned to bipyridine, phenanthroline and alkyne are highlighted in blue, yellow and purple, respectively.

Fig. 9. Transient absorption spectroscopy in acetonitrile.

Transient absorption spectra recorded for 1 (a) and 2 (c) in acetonitrile upon pumping at 400 nm at different delay times. Kinetic traces of 1 (Inset in a) and 2 (Inset in c) at selected wavelength. Decay-associated spectra and corresponding time constants of 1 (b) and 2 (d) derived from a global multiexponential fit applied on the transient absorption data.

Fig. 10. Transient absorption spectroscopy in dichloromethane.

Transient absorption spectra recorded for 1 (a) and 2 (c) in dichloromethane upon pumping at 400 nm at different delay times. Kinetic traces of 1 (Inset in a) and 2 (Inset in c) at selected wavelength. Decay-associated spectra and corresponding time constants of 1 (b) and 2 (d) derived from a global multiexponential fit applied on the transient absorption data (c).