Kinetics of ligand exchange in solution: a quantitative mass spectrometry approach

Complex speciation and exchange kinetics of labile ligands are critical parameters for understanding the reactivity of metal complexes in solution. We present a novel approach to determine ligand exchange parameters based on electrospray ionization mass spectrometry (ESI-MS). The introduction of isotopically labelled ligands to a solution of metal host and unlabelled ligands allows the quantitative investigation of the solution-phase equilibria. Furthermore, ion mobility separation can target individual isomers, such as ligands bound at specific sites. As a proof of concept, we investigate the solution equilibria of labile pyridine ligands coordinated in the cavity of macrocyclic porphyrin cage complexes bearing diamagnetic or paramagnetic metal centres. The effects of solvent, porphyrin coordination sphere, transition metal, and counterion on ligand dissociation are discussed. Rate constants and activation parameters for ligand dissociation in the solution can be derived from our ESI-MS approach, thereby providing mechanistic insights that are not easily obtained from traditional solution-phase techniques.


Time (s)
Relative intensity                .Typical ion transfer voltages were quadrupole ion energy = 3 eV and collision energy = 5 eV.The mass range scanned by the ToF analyzer was m/z 1200-2000.TIMS experiments were performed in N2 using the imeX Detect mode, by scanning ion mobility from 1.3 V.s.cm −2 to 2.08 V.s.cm −2 .The accumulation time was varied from 0 to 30 ms depending on the pyridine concentration, in order to maintain the maximum ion signal lower than 2.10 4 counts and thereby minimize ion activation in the ion mobility region.
The TIMS dimension was calibrated using five selected ions from the Agilent ESI LC/MS tuning mix [(1221.9906For experiments performed with the tBu4NCl salt, the End Plate Offset was varied from 0-80V, the nebulizer set to 0 Bar and the accumulation time was set to 50 ms to account for ion suppression.Computational details.DFT calculations were carried out at the B3LYP-D3/def2svp level with the Gaussian 16 package.The SMD model was used to account for implicit chloroform solvation.All reported structures correspond to potential energy surface minima as confirmed by analyses of the corresponding Hessian matrixes.Reported energies include zero-point vibrational energy corrections and thermal corrections calculated at the same level of theory. Reported ∆G values also include a correction of (1.9 Δn) kcal mol −1 to account for the change in number of moles (n) for reactions involving dissociation of a ligand.
Preparation of reaction mixtures.Reaction mixtures for DRL experiments were prepared by mixing stock solutions of the porphyrin complex and pyridine in a glass vial under continuous stirring to obtain a concentration of 4 µM of porphyrin complex, and 100 equiv.pyridine.The glass vial was placed in a home-made Peltier heater/cooler to control the reaction temperature (Figure S25), and the reaction mixture was infused into the ESI source of the timsToF instrument via a silica capillary by applying an overpressure of N2 (approx.2.5 psi).After two minutes, a minimal volume (80µL) of pyridine-D5 is added to the reaction mixture with a Hamilton syringe.Final concentrations (for addition of 100 equiv.pyridine and 100 equiv.pyridine-D5) : 3.82 µM of porphyrin compound, 380 µM pyridine, 380 µM pyridine-D5.X-ray crystallography.Reflections were measured on a Bruker D8 Quest diffractometer with sealed tube and Triumph monochromator (λ = 0.71073Å).Software package used for the intensity integration was Saint (v8.40a). 1 Absorption correction was performed with SADABS. 2 The structures were solved with direct methods using SHELXT-2014/5. 3 Leastsquares refinement was performed with SHELXL-2018/3 4 against | | of all reflections.Non-hydrogen atoms were refined freely with anisotropic displacement parameters.Hydrogen atoms were placed on calculated positions or located in difference Fourier maps.All calculated hydrogen atoms were refined with a riding model.

Crystal structure and structure refinement of CoC-Cl
Single crystals were grown by slow diffusion of n-heptane into a chloroform solution of the compound.equipped with an ESI ion source.UV-Vis spectra spectra were recorded at 298 K on a JASCO J-815 CD spectrophotometer (2 mm quartz cell).

Synthesis of CoC-Cl.
To an argon-purged solution of the corresponding free base porphyrin cage H2C (182 mg, 0.135 mmol) in chloroform (20 mL), methanol (20 mL) and triethylamine (1 mL) was added Co(OAc)2.4H2O(215 mg, 0.863 mmol).The mixture was heated at reflux under argon for 64 h.After cooling, the solvent was evaporated and the residue redissolved in CHCl3 (50 mL).The organic layer was subsequently extracted with brine (2 × 100 mL) and water (100 mL) and evaporated to dryness.The crude product was purified by column chromatography (silica 60H, 1% MeOH in CHCl3 (v/v)) and the first red fraction was collected and evaporated to dryness.The product was dissolved in a minimal amount of chloroform, the solution was filtered, and then added to rapidly stirred n-heptane (20 mL).
The formed precipitate was collected by centrifugation and dried under vacuum.It was subsequently dissolved in chloroform (5 mL), methanol (3 mL) and 12N aqueous HCl (2 mL) and the solution was stirred in air for 3 h.CHCl3 (50 mL) was added and the organic layer was extracted with water (2 × 100 mL) and evaporated to dryness.The product was dissolved in a minimal amount of chloroform, the solution was filtered, and then added to rapidly stirred n-heptane (20 mL).

Synthesis of CoC-PF6.
A suspension of CoC-Cl (34.0 mg, 23.6 mol) and AgPF6 (6.0 mg, 24 mol) in freshly distilled CH2Cl2 (2 mL) was stirred under argon in the dark for 16 h.Then, 5.0 mg of AgPF6 (19 mol) were added and stirring was continued for 2h.The suspension was filtered over a thin layer celite, which was washed with a mixture of CHCl3 and MeOH (9:1, v/v) until the washings remained colorless.The filtrate was evaporated to dryness and the crude product was purified by column chromatography (silica 60H, CHCl3/MeOH 9:1, v/v).The red fraction was collected and evaporated to dryness.The product was dissolved in a minimal amount of chloroform, the solution was filtered, and then added to rapidly stirred n-heptane (20 mL).The formed precipitate was collected by centrifugation and dried under vacuum.Yield: 18.0 mg (49%) of CoC-PF6 as a red solid.M.p. > 300 o C (dec).Synthesis of CoC*-Cl (S,S,S,S-enantiomer). Starting from the corresponding free base porphyrin cage H2C* (S,S,S,S-enantiomer, 19.0 mg, 0.0136 mmol) this compound was synthesized as described for CoC-Cl.Yield: 13.8 mg (70%) of a red solid.M.p. > 300 o C (dec). 1 H NMR (500 MHz, CDCl3) : the compound

Kinetic model
We built a kinetic model to describe the exchange of pyridine ligands bound at a particular site on the CoC complex in a DRL experiment, with X being any trans-axial ligand bound to the cage complex in solution: By Solving the kinetic model for small excess of ligand: The analytical solving of Equations S8 and S9 proved to be difficult.We thus decided to solve them numerically using Matlab's ode45 solver.The numerical simulations allow to evaluate the effect of systematic variations of the rate constants k1 and k-1, and initial concentrations of cage complex or pyridine(-D5).
However, finding initial conditions for the numerical solving of Equations DRL experiments performed with small excess of pyridine ligand and different delays (30s to 5 min -Figure S5) revealed that different delays between the addition of pyridine and pyridine-D5 do not produce significantly/statistically different data.The initial equilibrium between the CoC complex and pyridine ligands thus reaches steady state within 30s.Using the k-1 rate constant measured with large excess of ligand and the initial concentration of CoC complex, numerical solving of Equations S14, 15 and 16 shows that association rate constants larger than 50 000 M -1 s -1 are required to reach steady state within 30s (Figure S47).These experiments provide a lower range estimation for k1.Note that we used the k-1 value obtained by DRL for measurements on CoC-Cl with 200 equiv.pyridine(-D5), i.e., 5.4*10 -4 s -1 .We then varied systematically one parameter (either k1, k-1 or initial concentration of pyridine(-D5)), while keeping the other ones constant, and monitored the effect on kMS.As shown in Figure S49, our model predicts a decrease of kMS with increasing initial concentrations of (pyridine + pyridine-D5).At large excess of (pyridine + pyridine-D5), kMS becomes equal to k-1, in agreement with the approximation discussed in the previous paragraph.With large excess of pyridine, kMS is always equal to k-1 as evidenced by Figure S50, in which we varied k-1 while keeping k1 and initial concentrations of (pyridine + pyridine-D5) constant.k 1 = 100 000 M -1 s -1 k 1 = 500 000 M -1 s -1 k 1 = 1 000 000 M -1 s -1 C 0 CoC = 4 µM k -1 = 0.00054 s -1 Figure S51.Evolution of kMS with increasing initial concentrations of (pyridine + pyridine-D5) from the numerical solving of the kinetic model.The rate constant k-1 was kept at 0.00054 s -1 and k1 was varied from 50 000 M -1 s -1 to 1 000 000 M -1 s - Figure S1.a) Evolution of kMS as measured by DRL experiments in CHCl3 at 24°C with increasing equivalents of (pyridine + pyridine-D5).For each experiment, a 1:1 ratio of pyridine and pyridine-D5 was used.The green line is a guide for the eyes.b) Relative time evolution of [CoC(Pyridine)] + (red dots) and [CoC(Pyridine-D5)] + (blue dots) in DRL experiments recorded at 24°C with CoC-Cl after addition of various equiv. of pyridine(-D5) in CHCl3.Dots are experimental data points and lines correspond to fittings of experimental data by Equation S1 (red lines) and S2 (blue lines).
Figure S5.a) Relative time evolution of [CoC(Pyridine)] + (red dots) and [CoC(Pyridine-D5)] + (blue dots) in DRL experiments recorded at 24°C with CoC-Cl after addition of 1.25 equiv. of pyridine and 1.25 equiv.pyridine-D5 in CHCl3, with different delays.Dots are experimental data points and red lines correspond to fittings of experimental data by Equation S1.Each experiment was repeated five times, and each replica is denoted by R1, R2, R3, R4 and R5.b) Average values of kMS measured for each time delay between addition of pyridine and pyridine-D5.Error bars correspond to the standard deviation.

Figure S6 .Figure S7 .
Figure S6.Relative time evolution of [CoC(Pyridine)] + (red dots) and [CoC(Pyridine-D5)] + (blue dots) in DRL experiments recorded at 18°C, 24°C, 30°C, 35°C and 40°C for CoC-Cl after addition of 100 equiv.pyridine and 100 equiv.pyridine-D5 in CHCl3.Dots are experimental data points and red lines correspond to fittings of experimental data by Equation S1.Each experiment was repeated three times, and each replica is denoted by R1, R2, R3.The values of kMS and [CoC + Py]eq obtained from fittings are indicated for each experiment.

Figure S8 .
Figure S8.Free energy scheme for the pyridine ligand dissociation mechanism for [Co III C(ClOUT)(PyIN)].Spin states of the Co(III) are written as follow: Singlet 1 CoC (black), Triplet 3 CoC (blue), Quintet 5 CoC (red).DFT structures are only shown for the 1 CoC state.H atoms are omitted for clarity.

Figure S9 .
Figure S9.Free energy scheme for the inner pyridine ligand dissociation mechanism for [Co III C(PyOUT)(PyIN)].Spin states of the Co(III) are written as follow: Singlet 1 CoC (black), Triplet 3 CoC (blue), Quintet 5 CoC (red).DFT structures are only shown for the 1 CoC state.H atoms are omitted for clarity.

Figure S10 .
Figure S10.Relative time evolution of [CoC(Pyridine)] + (red dots) and [CoC(Pyridine-D5)] + (blue dots) in DRL experiments recorded at 24°C, 30°C, 40°C and 50°C for CoC-Cl after addition of 100 equiv.pyridine and 100 equiv.pyridine-D5 in acetonitrile.Dots are experimental data points and red lines correspond to fittings of experimental data by Equation S1.Each experiment was repeated three times, and each replica is denoted by R1, R2, R3.The values of kMS and [CoC + Py]eq obtained from fittings are indicated for each experiment.

Figure S14 .Figure S15 .
Figure S14.Relative time evolution of [CoC*(Pyridine)] + (red dots) and [CoC*( Pyridine-D5)] + (blue dots) in DRL experiments recorded at 18°C, 24°C, 30°C, 35°C and 40°C for CoC*-Cl after addition of 100 equiv.pyridine and 100 equiv.pyridine-D5 in CHCl3.Dots are experimental data points and red lines correspond to fittings of experimental data by Equation S1.Each experiment was repeated three times, and each replica is denoted by R1, R2, R3.The values of kMS and [CoC* + Py]eq obtained from fittings are indicated for each experiment.

Figure S16 .Figure S17 .Figure S18 .Figure S19 .
Figure S16.Relative time evolution of [CoC(Pyridine] + (red dots) and [CoC(Pyridine-D5)] + (blue dots) in DRL experiments recorded at 18°C, 24°C, 30°C, 35°C and 40°C for CoC-PF6 after addition of 100 equiv.pyridine and 100 equiv.pyridine-D5 in CHCl3.Dots are experimental data points and red lines correspond to fittings of experimental data by Equation S1.Each experiment was repeated three times, and each replica is denoted by R1, R2, R3.The values of kMS and [CoC + Py]eq obtained from fittings are indicated for each experiment.

Figure S20 .
Figure S20.DFT calculated structures of [MnC(ClOUT)(PyIN)] and [CoC(ClOUT)(PyIN)] in their most stable spin state, i.e., quintet and singlet respectively, optimized with B3LYP-D3/def2svp.The distance between the metal centre and the nitrogen atom of the pyridine is highlighted.Hydrogen atoms are not shown for clarity.

Figure S21 .
Figure S21.Free energy scheme for the inner pyridine ligand dissociation for [Mn III C(PyIN)(ClOUT)].Spin states of the Mn(III) are written as follow: Singlet 1 MnC (black), Triplet 3 MnC (blue), Quintet 5 MnC (red).DFT structures are only shown for the 5 MnC state.H atoms are omitted for clarity.

Figure S23 .
Figure S23.Free energy scheme for the outer pyridine ligand dissociation mechanism for [Co III C(PyOUT)(PyIN)].Spin states of the Co(III) are written as follow: Singlet 1 CoC (black), Triplet 3 CoC (blue), Quintet 5 CoC (red).DFT structures are only shown for the 1 CoC state.H atoms are omitted for clarity.Note that the calculations of the triplet and quintet states of [CoC(PyIN)] + did not converge, and are not shown.

Figure S25 .
Figure S25.Picture of the experimental setup in which the DRL reaction mixture is contained in the glass vial placed intro a home-made Peltier heater/cooler.A flow of N2 is applied to the reaction mixture to infuse the solution from the vial to the ESI source through a silica capillary highlighted by red stripes.

1 Figure S49 .
Figure S49.Evolution of kMS with increasing initial concentrations of (pyridine + pyridine-D5) from the numerical solving of the kinetic model.The rate constants k1 and k-1 were kept at 50 000 M -1 s -1 and 0.00054 s -1 respectively.

1 Figure S50 .
Figure S50.Evolution of kMS with k-1 from the numerical solving of the kinetic model.The rate constantk1 was kept at 50 000 M -1 s -1 and 100 equiv. of pyridine and pyridine-D5 were considered, relative to the initial concentration of CoC.
The formed precipitate was collected by centrifugation and dried under vacuum.Yield: 122 mg (63%) of CoC-Cl as a red solid.Single crystals suitable for X-ray analysis were grown by slow diffusion of n-heptane in a CDCl3 solution of the compound.M.p. > 300 o C (dec).
Figure S47.Concentration of [CoC(Py)], [Py] and [CoC] obtained by numerical solving of Equations S14, 15 and 16 with variable k1 and fixed values of k-1 = 0.00054 s -1 , CCoC 0 = 4 µM and 1.25 equiv.pyridine relative to initial CoC concentration.We used the steady-state concentrations of [P] and [CoC] as initial conditions for the numerical solving of Equations S8 and S9.The concentrations of [P], [Y] obtained from the solver are then renormalized in the same way as we renormalize the ion currents of [CoC(Py)] + and [CoC(Py L )] + (See Figure S48).The normalized concentrations are then fitted using Equations S1 and S2, thereby affording kMS. 1.