The Mechanism of Rh(I)-Catalyzed Coupling of Benzotriazoles and Allenes Revisited: Substrate Inhibition, Proton Shuttling, and the Role of Cationic vs Neutral Species

Direct coupling of benzotriazole to unsaturated substrates such as allenes represents an atom-efficient method for the construction of biologically and pharmaceutically interesting functional structures. In this work, the mechanism of the N2-selective Rh complex-catalyzed coupling of benzotriazoles to allenes was investigated in depth using a combination of experimental and theoretical techniques. Substrate coordination, inhibition, and catalyst deactivation was probed in reactions of the neutral and cationic catalyst precursors [Rh(μ-Cl)(DPEPhos)]2 and [Rh(DPEPhos)(MeOH)2]+ with benzotriazole and allene, giving coordination, or coupling of the substrates. Formation of a rhodacycle, formed by unprecedented 1,2-coupling of allenes, is responsible for catalyst deactivation. Experimental and computational data suggest that cationic species, formed either by abstraction of the chloride ligand or used directly, are relevant for catalysis. Isomerization of benzotriazole and cleavage of its N–H bond are suggested to occur by counteranion-assisted proton shuttling. This contrasts with a previously proposed scenario in which oxidative N–H addition at Rh is one of the key steps. Based on the mechanistic analysis, the catalytic coupling reaction could be optimized, leading to lower reaction temperature and shorter reaction times compared to the literature.


[Rh(DPEPhos)(BTAH)2]BF4 (3)
[Rh(DPEPhos)(COD)]BF4 (8.4 mg, 0.01 mmol) was dissolved in 2 mL of MeOH and hydrogenated for 33 seconds due to its pre-hydrogenation time.Hydrogen was removed in three freeze-pump-thaw cycles.Benzotriazole (2.7 mg, 0.023 mmol) was added to the solution and the mixture was stirred for one hour.The solvent was removed in vacuum and the residue was dissolved in 0.6 mL of THF and layered with 5 mL of Et2O.After a couple of days, orange crystals could be isolated and analyzed by X-ray single crystal analysis.[Rh(DPEPhos)(COD)]BF4 (83.6 mg, 0.10 mmol) was dissolved in MeOH (10 mL) and the orange suspension was hydrogenated for 4 min.The hydrogen was removed from the dark red solution in three freeze-pump-thaw cycles.Cyclohexyl allene (29 µL, 0.2 mmol) was added and the solution was stirred for 90 min and additionally dried in vacuum.The residue was dissolved in 2.5 mL of MeOH and stored at 4 °C.Overnight an off-white precipitate formed, that was separated from the orange solution.The product was washed three times with Et2O (ca. 5 mL) and finally dried in vacuum.Circa 43 mg (0.05 mmol) of product was isolated (49% yield).Colourless needles suitable for X-ray crystallography were grown from a saturated solution in MeOH.

High field monitoring of the complete catalytic reaction
[Rh(µ-Cl)(DPEPhos)]2 (9.5 mg, 0.007 mmol) was dissolved in 1,2-DCE-d4 (600 µL) and stirred for 30 minutes at room temperature.Cyclohexyl allene (60 µL, 0.4 mmol) was added to the precatalystbenzotriazole mixture and the mixture was stirred for 30 seconds at room temperature.The solution was transferred to a Young's tap NMR tube, in which BTAH (36.0 mg, 0.302 mmol) was placed.The measurements were performed at a reaction temperature of 50 °C.

Low field monitoring of the complete catalytic reaction
[Rh(µ-Cl)(DPEPhos)]2 (9.5 mg, 0.007 mmol) and BTAH (36.0 mg, 0.302 mmol) were dissolved in                      (SHELXT) 5 or direct methods (SHELXS-97) 6 and refined by full matrix least square techniques against F 2 (SHELXL-2014) 7 .Semi-empirical absorption corrections were applied (SADABS/Bruker). 8e non-hydrogen atoms were refined anisotropically.The hydrogen atoms, except for the hydrogens at the nitrogen atoms of 2 (determination and refinement from electron density), were placed in the theoretical positions and were refined by using the riding model.
Contributions of solvent molecules were removed in 2 and 6 from the diffraction data with PLATON/SQUEEZE. 9DIAMOND (Crystal Impact GbR) was used for structure representations.

TD-UV-vis computation
The UV-vis spectrum of complex 5 in MeOH shows no absorption maxima in the visible region of the spectrum (Figure S41).Assignment of the transitions was done using time-dependent DFT (TD-DFT) UV-vis computations.
TD-DFT calculations were performed to predict UV-vis absorption spectra on Gaussian 16 using the B3LYP method, a def2-SVPP basis set and a PCM solvent model for methanol considering 40 excited states.The geometry of the unoptimized molecular structure was used for this calculation.The calculated data were plotted with a half-width of 2000 cm -1 and normalized to the maximum.The corresponding simulated UV-vis spectrum and the oscillator strength of the excited states is shown in Figure S40 and the first 20 excited states are reported in the following.
The calculated and experimental spectra are in very good agreement, with the calculated absorption maximum at 265 nm.According to the charge density difference analysis of the first exited states, all absorption bands correspond to metal-to-ligand charge transitions together with prominent ligand-toligand charge transfer.

Figure S39 .
Figure S39.Contour plots of the Laplace operators of the electron density ∇ 2 r of 5 in 3 different planes; a) C47-Rh1-C58, b) C48-Rh1-C50, c) C47-C48-C58.Dashed lines show negative values (local charge concentration), solid lines show positive values (local charge depletion).The Laplace plot is overlaid with the molecular diagram from the QT-AIM analysis and the Wiberg binding indices (italics).Brown lines indicate bond paths, blue dots correspond to critical bond points, and orange dots indicate critical ring points.Geometry optimized by BP86+GD3/def2-SVPP.
Crystallographic data (excluding structure factors) for the structures reported in this paper have been deposited at the Cambridge Crystallographic Data Centre.Copies of the data can be obtained free of charge on application to CCDC, 12 Union Road, Cambridge, CB21EZ, UK (fax: int.code + (1223) 336-033; e-mail: deposit@ccdc.cam.ac.uk

Table S2 .
Optimized structures of 5 with different methods and a def2-SVPP basis set and the deviation from the geometry of the molecular structure.

Table S3 .
Optimized structures of 5 with different methods, basis sets and empirical dispersion and the deviation from the geometry of the molecular structure.

Table S4 .
Optimized structures of 5 with B3LYP/def2-SVPP and GFN2-xTB and the deviation from the geometry of the molecular structure.

Table S5 .
Orbital representation of selected excited states of the TD-UV-vis computation of 5.