EPR Study of the Low-Spin [d³; S =1/2], Jahn-Teller-Active, Dinitrogen Complex of a Molybdenum Trisamidoamine

Abstract

The first EPR study of the trisamidoamine complex, [Mo]N₂, where [Mo] = [Mo(III)[HIPTN₃N]^3- = [3,5-(2,4,6-i-Pr₃-C₆H₂)₂C₆H₃NCH₂CH₂]₃N^3-], reveals that this low-spin (S = ½) [d³] complex exhibits a ²E state that undergoes a pseudo Jahn-Teller distortion in the adiabatic limit, modified by interactions with the solvent, and gives approximate values of interaction energies. The experiments establish that [Mo]N₂ exhibits the low-spin [e³] electronic configuration, not [a²e¹], with the a(z²) antibonding orbital substantially higher in energy than the e[xz, yz] orbitals.

The sterically encumbered trisamidoamine (TAA) complex^1,2 [Mo]=Mo[HIPTN₃N], Inset 1, is extremely important for its ability to catalyze the reduction of N₂ to NH₃,³ and for the mechanistic insights into this process that it provides; of the thirteen compounds that are postulated to be involved in the formation of 2NH₃, the structures of eight already have been determined by x-ray diffraction.^4,5 Nonetheless, at present there is little understanding of the particular features of the electronic structure of the [Mo]L complexes that underlie their reactivity.

Inset 1.

The key first step in the catalytic cycle is N₂-bound [Mo]N₂,⁶ (Inset 1) which contains a Mo(III) that has been shown by magnetic susceptibility measurements to have a low-spin (ls) [d³; S=1/2] electronic configuration. It has not been emphasized that this is one of the rarest configurations in coordination chemistry. To our knowledge, there is only one report of a ls-[d³] complex of any transition ion prior to the synthesis of the TAA complexes.⁷ Nor has it been noted that a ls-[d³] ion in the trigonal-bipyramidal coordination environment of [Mo]N₂ exhibits a doubly-degenerate ²E ground state, Scheme 1, that is subject to a Jahn-Teller (JT) distortion by vibronic coupling to doubly-degenerate e₂ vibrations.^8,9 Indeed, as shown in the scheme, this ls-[d³] complex has two possible electronic configurations: [e³] and [a²e¹], each doubly degenerate. Although the former is suggested to apply,⁴ this has not been tested. This report first describes the possible vibronic behaviors of [Mo]N₂. It then presents an EPR study that characterizes the electronic configuration and vibronic properties of this complex.

Scheme 1.

Both [d³; S=1/2] configurations of Scheme 1 exhibit unquenched orbital angular momentum, and splitting of the orbital degeneracy by spin-orbit coupling (SOC) competes with the JT effect (JTE). Linear vibronic coupling is describable in terms of a single composite (‘interaction’) e₂ mode,⁸ in which the molecule undergoes a ‘pseudo-Jahn-Teller’ (PJT) distortion, ρ₀,⁸

$${\rho }_0={[{(\text{F}/\text{K})}^2-{(\lambda /2\text{F})}^2]}^{1/2}$$ (1)

provided the JT vibronic coupling is strong enough, namely ρ₀ is real. Here K is the effective force constant of the interaction mode, F is the linear coefficient of vibronic coupling to this mode, and λ is the SOC constant. This JTE replaces the electronic degeneracy with a vibronic degeneracy in which the complex is distorted (e.g.: equilateral → isosceles triangle; axial → non-axial N₂) and the distortion ‘pseudo-rotates’. If the JT vibronic coupling is weak (ρ₀ is imaginary) the JT distortion is quenched and [Mo]N₂ retains the trigonal symmetry.

The molecular adiabatic potential energy surface (APES) of an isolated molecule is a function of the PJT distortion, ρ, but is independent of the phase angle (φ) that defines the direction of the distortion. One must further allow for a solvent-dependent environmental contribution to the e-orbital splitting, the ‘solvent potential’, V_L,¹⁰ resulting in APES energies, W_± = Kρ²/2±2[(λ/2)² + (Fρ sinφ)² + (Fρ cosφ) + V_L)²]^1/2. When V_L = 0, the ground APES (W_-) is the well-known ‘modified Mexican hat’, Fig 1A, with the minimum of the ‘trough’ at ρ₀.^8,9 V_L ≠ 0 skews the ground APES, Fig 1B, favoring a particular distortion. At low temperatures, the dynamic PJT distortion will localize at the skewed APES minimum; localization also is favored by quadratic JT coupling.⁸

Figure 1.

Lower PJT APES (W_-) calculated with (A) λ = 800 cm ⁻¹, K = 1 mdyn/A, Fρ₀ = 0.95λ; (B) as in A, but with V_L = 0.1λ.

The X-band CW EPR spectrum of S = ½ [Mo]N₂ in toluene, Fig 2, exhibits axial symmetry, with g_⊥ = 1.61 < 2 < g_‖ = 3.03; identical values were obtained from Q-band CW and spin echo-EPR spectra.^11,12 The g-values distinguish unambiguously between the [e³] and [a²e¹] configurations, and give insights into the vibronic and solvent interactions. We begin by considering a two-orbital model Hamiltonian in which vibronic coupling with the interaction mode combines with SOC to mix and split the e(xz, yz; m_S = ±1) orbitals.¹³ This PJT model gives an axial g-tensor for the ground APES with g_‖ along the trigonal axis. The [g_‖, g_⊥] components at the APES minimum depend on the ratio of the sum of equilibrium distortion and environmental energies to the SOC, [(Fρ₀) + V_L/2]/λ, through the fictitious angle 2θ;

Figure 2.

9.36 GHz EPR of [Mo]N₂. The ‘gaps’ at g ∼ 2 in some spectra eliminate signals from decomposition products and/or precipitated complex in the poorly-glassing THF and octane, chosen to vary V_L.

$$tan2\theta =2[{\text{F}\rho }_0+{\text{V}}_{\text{L}}/2]/\lambda .$$ (2)

$${[\text{e}}^{\text{3}}]{\text{g}}_{\parallel }=2(1+kcos2\theta )$$ (3a)

$$[{\text{a}}^2{\text{e}}^1]{\text{g}}_{\parallel }=2(1-kcos2\theta )$$ (3b)

$${\text{g}}_{\perp }=(2sin2\theta )$$ (3c)

the quantity (1 - k) corresponds to d-electron delocalization. As 2θ increases from 0 (JTE quenched; no distortion) to π/2 (SOC quenched), the [e³] configuration gives [ 0 ≤ g_⊥ ≤ 2; 4 ≥ g_‖ ≥ 2], while [a²e¹] gives [0 ≤ g_‖ ≤ g_⊥ ≤ 2]. The observed g-values thus require the [e³] configuration for [Mo]N₂. Analysis further gives d-electron delocalization of (1-k) = 15% and a distortion energy with tan 2θ = 1.4. The PJT splitting between the lower APES minimum and the upper APES is large, ΔE(ρ₀) = 2λ[(1/2)² + ((tan 2θ)/2)²]^1/2 ∼ 2λ ∼ 1600 cm^-1.¹⁴ As a result, the ground and excited APES should be vibronically decoupled (‘strong-coupling’ or adiabatic limit), and the system well-described by the ground APES and its g-values.^8,9 The JT energies may even be large enough to have an influence on the energetics of reaction.

Variations in the g-values with solvent, g_‖ = 3.03 (toluene), 3.10 (THF), 3.26 (octane) (Fig 2) reveal the presence of an environmental influence.¹⁰ Taking octane as approximating a non-perturbing environment, V_L(octane) → 0, the g-values provide interaction parameters through eq 2: Fρ₀/λ ∼ 1; V_L(THF)/ λ ∼ 0.3, and V_L(toluene)/λ ∼ 0.4. The small environmental term, Fρ₀ > V_L, contrasts with the PJT-active MCp₂,^10,16 M = Co(II), Fe(I), and Fe(III), where Fρ₀ ≪ V_L,.^10,15-17 The reversal of this inequality for [Mo]N₂ is noteworthy, as the bulky HIPT substituents were in fact incorporated to shield the metal center from its environment.

The g-values of [Mo]N₂ in toluene do not shift significantly as temperature increases, whereas in THF and octane, solvents with smaller V_L, the g-values shift slightly (Fig 2). Unlike the well-studied JTE for Cu(II),^8,9,17 this behavior cannot be used as evidence regarding possible reorientation of a trapped JT distortion. Warping of the ground APES of Cu(II) ions leads to three alternate JT distortion orientations with different g-tensor orientations, and thermally activated reorientation of the distortion causes strong ‘motional’ effects in the Cu(II) g-values. In contrast, g_‖ for [Mo]N₂ is oriented along the trigonal axis for all orientations of the PJT distortion (angles, φ), so reorientation has no effect. We instead attribute g-shifts to thermal excitations within the energy ‘well’ of the skewed APES (Fig 1B), which is shallower for smaller values of V_L. The EPR signal for [Mo]N₂ in toluene does not broaden with increasing temperature; those for [Mo]N₂ in THF and octane broaden only slightly (Fig 2). This contrasts with the substantial broadening from Orbach relaxation by low-lying states of the PJT-active MCp₂,^10,15 and indicates that [Mo]N₂ has no low-lying excitations (δ, Scheme 1, is large).

In summary, this first EPR study of an [Mo]L complex establishes that [Mo]N₂ exhibits the ls [e³] electronic configuration and the JTE in the vibronic strong-coupling (adiabatic) limit. The anti-bonding a(z²) orbital is considerably higher in energy than e(xz, yz), presumably because of interactions with the axial nitrogen atoms. It appears that interactions with the solvent localize the PJT distortion of [Mo]N₂ at low temperatures, even though the complex is highly sterically encumbered. If there is no PJT distortion or a pseudo-rotating one, hyperfine tensors of the in-plane N atoms as determined by ¹⁴,¹⁵N ENDOR studies of [Mo]N₂ (in progress) will be identical; if skewing of the APES selects a particular distortion, as proposed, they will not be equal. DFT computations also are in progress to identify the vibrations that contribute to the composite interaction mode. ENDOR studies of [Mo]N₂ and other [Mo]L also will permit comparison with intermediates trapped during the turnover of nitrogenase.^19,20