The importance of ligand-induced backdonation in the

stabilization of square planar d10 Nickel π-complexes

Addison N. Desnoyer,† Weiying He,† Shirin Behyan, Weiling Chiu,
Jennifer A. Love,* and Pierre Kennepohl*[a]
Abstract: The electronic nature of Ni π-complexes is underexplored “oxidative addition” reaction to occur at a Zr(IV), d0 metal centre.[6]
even though they are widely postulated as intermediates in In another example, we have recently shown that the oxidation of
organometallic chemistry. Herein, we probe the geometric and (TPA)Rh olefin complexes (TPA = tris(2-pyridylmethyl)amine)
electronic structure of a series of nickel π-complexes, Ni(dtbpe)(X) with H2O2 to form 2-rhodaoxetanes[7] is more accurately described
(dtbpe = 1,2-bis(di-tert-butyl)phosphinoethane; X = alkene or carbonyl as a ligand-centred oxidation,[8] rather than a metal-centred
containing π-ligands), using a combination of 31P NMR, Ni K-edge oxidation.[9] In this case, the π-ligand acts as a two-electron redox
XAS, Ni Kβ XES, and DFT calculations. These complexes are best centre.
described as square planar d10 complexes with π-backbonding acting We have recently become interested in exploring the fundamental
as the dominant contributor to M-L bonding to the π-ligand. The organometallic chemistry of earth-abundant, first-row transition
degree of backbonding correlates with 2JPP from NMR and the energy metals. For example, we are exploring the organometallic
of the Ni 1s→4pz pre-edge in the Ni K-edge XAS data, and is chemistry of nickel,[10–14] which has undergone a renaissance in
determined by the energy of the π*ip ligand acceptor orbital. Thus, recent years.[15–21] Our focus has been the structure and reactivity
unactivated olefinic ligands tend to be poor π-acids whereas ketones, of nickel π-complexes, which have been reported in a wide range
aldehydes, and esters allow for greater backbonding. However, of catalytic processes, including the coupling of CO2 and
backbonding is still significant even in cases where metal ethylene,[5,22–32] intermolecular Tischenko coupling,[33–35]
[36,37]
contributions are minor. In such cases, backbonding is dominated by benzoxasilole synthesis, the aldol reaction,[38] allylic
charge donation from the diphosphine, which allows for strong alkylation,[39] allylic amination,[40] allylic amidation,[41] epoxide
backdonation even though the metal centre retains a formal d10 functionalization,[42] and Suzuki-Miyaura coupling.[43] Nickel π-
electronic configuration. This ligand-induced backbonding can be complexes of heteroarenes have also been identified as key
formally described as a 3-centre-4-electron (3c-4e) interaction where intermediates in nickel-catalysed catalyst transfer
the nickel centre mediates charge transfer from the phosphine σ- polycondensation to form polythiophenes.[44–56] Given the
donors to the π*ip ligand acceptor orbital. The implications of this importance of nickel π-complexes, detailed exploration of their
bonding motif are described with respect to both geometric structure structure and reactivity is needed. Herein, we report the electronic
and reactivity. structures of a series of nickel π-complexes relevant to catalysis.
Additionally, we identify the impact of ancillary ligands in inducing
and supporting π-backbonding, even in cases where metal
contributions are limited.
Introduction In previous work, we noted that the 31P{1H} NMR spectroscopic
data of a number of (dtbpe)Ni (dtbpe = 1,2-bis(di-tert-
Over the last two decades, renewed interest in the redox non- butyl)phosphino)ethane) π-complexes were consistent with
innocence of ligands has led to their proliferation in inorganic typical d10 Ni(0) complexes (Chart 1).[10,11,57] In contrast, we also
chemistry.[1,2] The use of these ligands as electron reservoirs noted that the distorted square planar geometry with significant
enables two-electron processes from complexes which typically elongation of the π-bond were most consistent with a d8 Ni(II)
exhibit single-electron chemistry, particularly first-row transition formulation, in keeping with the metallaepoxide extreme of the
metals.[3] In a pioneering example, Chirik and co-workers Dewar-Chatt-Duncanson (DCD) model of bonding (Scheme 1). In
demonstrated that (PDI)Fe(N2)2 (PDI = 2,6-(2,6- addition, preliminary density functional theory (DFT) calculations
i
Pr2C6H3NCR)2C5H3N, R = Me or Ph) catalyses the formal [2+2] (vide infra) revealed prohibitively high barriers to rotation of the π-
cyclization of diolefins to form cyclobutane rings.[4] Notably, the ligand (i.e. 80-100 kJ/mol), demonstrating that these complexes
iron centre stays in the Fe(II) oxidation state throughout the have a strong preference for the square planar geometry despite
catalytic cycle, with the PDI ligand acting as a two-electron steric constraints. Experimentally, the 31P{1H} NMR spectrum of 1
reservoir. More recently, the Tsurugi, Arnold, and Mashima up to 110 °C reveals no dynamical processes, indicating that the
groups reported that both the geometric and electronic non- barrier to carbonyl inversion is greater than 70 kJ/mol. We have
innocence of α-diimine ligands plays a key role in niobium- previously reported similar high barriers to rotation with a
catalysed chlorination of olefins, where the metal centre stays in rhodium-olefin system.[8] Indeed, the metallaepoxide electromer
the Nb(V) oxidation state and redox events occur on the diimine of related nickel complexes have recently been invoked by the
ligand.[5] groups of Doyle[42] and Ogoshi[37] based on reactivity studies, and
In contrast to these open-shell systems, closed-shell systems that is also shown explicitly in Group 4 complexes that display similar
rely on the ligand accepting and/or donating electron pairs are structural parameters to the nickel species discussed here.[58–63]
less common. A notable example of this type of reactivity is the Ambiguity in the electronic structure of these nickel π-complexes
zirconium system reported by Heyduk, which allows for a putative hinders the effort towards rational design of nickel-catalysed
processes. We thus set out to investigate the bonding and
† electronic structure of a family of (dtbpe)Ni complexes by utilizing
these authors contributed equally to this work
[a] Department of Chemistry, The University of British Columbia, 2036 a combination of spectroscopic and computational techniques.
Main Mall, Vancouver, BC, Canada V6T 1Z1
E-mail: jenlove@chem.ubc.ca, pierre@chem.ubc.ca
Chart 1 List of π-complexes considered in this study identified by their Alternatively, differences in the geometry of four-coordinate
compound number (bold), τ4 value, sums of the angles about the metal centre complexes can be evaluated using τ4 values, which range from
(∑ ∡𝑁𝑁𝑁𝑁 ) and carbons in the π-ligand (∑ ∡𝐶𝐶 ), C=X bond distance of the π-ligand (𝐷𝐷 ) (𝑇𝑇 )
𝜏𝜏4 4ℎ = 0 to 𝜏𝜏4 𝑑𝑑 = 1.[67] This approach confirms the pseudo-
(𝑟𝑟𝐶𝐶𝐶𝐶 ), and NMR P,P coupling constants (𝐽𝐽𝑃𝑃𝑃𝑃 )..
tetrahedral (~𝑇𝑇𝒅𝒅 ) geometry of 13-15 (Chart 3), but suggests that
the more planar complexes split into a set of highly symmetrical
square planar complexes (Chart 2) and a set of complexes that
deviate more strongly from idealized D4h symmetry (Chart 1). The
latter complexes are all π-complexes where the deviation from an
idealized geometry results from the extremely small bite angle
formed by the π-ligand (when considered as an η2 ligand), even
while maintaining planarity. The planar geometry at the metal
centre implies that 1-7 exhibit a large degree of backbonding,
which would typically be ascribed to the formation of square
planar Ni(II) d8 complexes (i.e., a metallocyclic electronic
configuration as depicted in Scheme 1).

Scheme 1 Continuum of possible electronic configurations for binding of a π-

system to a redox-active metal centre. On the left, is the limiting case of simple
π-adduct formation, where M-L binding occurs via σ donation from the π-
system. As π-backbonding increases, the X=Y π bond weakens and, in the limit,
a metallacycle is formed with loss of the π bond and formal 2e- oxidation at the
metal centre.

The structure of the π-ligand itself has also frequently been used
to estimate the degree of backbonding: electron donation into the
ligand π* orbital via backbonding should lead to bond elongation.
For example, Zeise’s salt [KPtCl3(C2H4)] and Cramer’s dimer
[Rh(C2H4)2Cl]2, both commonly used organometallic starting
materials, feature short C=C ethylene bond distances of 137.5
This study is also relevant to the ongoing discussion about the
pm[68] and 139.5 pm[69], respectively. In contrast,
value of formal oxidation states.[64–66] Overall, we have found that
(MeTPA)Rh(C2H4)(BPh4), features a much longer C=C bond
these systems are dominated by π-backbonding with minimal 𝜎𝜎-
distance of 145 pm,[70] which corresponds to the
donation from the π-acidic ligand; the degree of backbonding
metallacyclopropane end of the DCD spectrum. However, this
reflects the π acidity of the ligand as well as the ability of the
method is generally qualitative, with many examples that fall in
ancillary diphosphine ligands to induce π-backbonding mediated
the middle of the spectrum being simply described as hybrids of
through the nickel 3d orbitals. We believe that this insight will
the two resonance forms.[71,72] Indeed, the C=O bond lengths of
prove beneficial to both the logical improvement of known
the η2-carbonyl complexes examined here (complexes 1, 2, 3 and
catalytic reactions with nickel and to the rational design of new
6) all fall between 131.7-135.4 pm.[73–75] This range is
transformations.
unfortunately ambiguous, as it is in the middle of the typical bond
lengths of ~122 pm and ~143 pm for C=O double bonds and C-O
single bonds, respectively. Similarly, information about the degree
Results and Discussion of backbonding can be gleaned from the sum of the bond angles
about the carbon atom of the π-unit (∑ ∡𝐶𝐶 , Chart 1). However,
Solid-state molecular structures these results are again inconclusive, as the observed ∑ ∡𝐶𝐶 (=
We selected a variety of (dtbpe)Ni complexes, ranging from well- 341 − 352°) are intermediate between those expected for planar
defined nickel(II) complexes to π-complexes of organic molecules sp2-hybridized and pyramidal sp3-hybridized carbon atoms. This
(Charts 1-3). The complexes were split into two categories based approach also suffers from the fact that many π-ligands bear
on the dihedral angles observed in the solid-state structures: (i) hydrogen substituents, which can be difficult to locate using
those with near planar geometries (1-12, where𝜑𝜑𝑑𝑑𝑑𝑑ℎ ~ 0° and traditional X-ray diffraction (XRD)[76] and occasionally require
∑ ∡𝑁𝑁𝑁𝑁 ~360°) and those with pseudo-tetrahedral geometries at the neutron diffraction experiments to accurately ascertain their
nickel centre (13-15, where 𝜑𝜑𝑑𝑑𝑑𝑑ℎ ~90° and ∑ ∡𝑁𝑁𝑁𝑁 ~440°). positions.
Chart 2. List of reference Ni(II) square planar complexes, which exhibit both electron density at the metal centre. We thus turned to X-ray
small 2𝐽𝐽𝑃𝑃𝑃𝑃 coupling constants (where available) and τ4 values. absorption spectroscopy (XAS) and X-ray emission spectroscopy
(XES) for an independent evaluation of the electronic structure
and spectroscopic oxidation states at the metal. We sought to
probe how a more direct measurement of 𝑍𝑍𝑒𝑒𝑒𝑒𝑒𝑒 at the metal centre
correlates to the above-discussed methods, in particular NMR
spectroscopy. To the best of our knowledge, a study relating the
magnitude of the NMR data (especially 31P coupling constants)
and how they correlate with alternate spectroscopic approaches
has not been performed.

Ni K-edge X-ray absorption spectroscopy

In principle, the spectroscopic oxidation state of the metal ion can
be accurately assigned via the energy of either the ionization edge
and/or the low-energy pre-edge features in the metal K-edge
near-edge spectrum.[86–92] Ni K-edge XAS can therefore be used
to explore the spectroscopic oxidation state of a wide range of
nickel-containing species.[93–97]

Chart 3. List of reference pseudo-tetrahedral Ni(0) complexes, with large τ4

values and ∑ ∡𝑁𝑁𝑁𝑁 > 400°.

Figure 1. Normalized Ni K-edge PFY XANES edge spectra for Ni(dtbpe)Cl2

(12), [Ni(dtbpe)]2(benzene) (7), Ni(dtbpe)(ethylene) (4), Ni(dtbpe)CF3COOEt
(1), and Ni(dtbpe)CF3COSEt (2). The pre-edge region for each of the spectra is
shown in the inset with assignments for the observed features.

Ni K-edge XAS data was obtained for several Ni π-complexes, in

addition to several reference complexes from Chart 1. Near-edge
Nuclear magnetic resonance spectroscopy spectra for ‘classic’ Ni(II) complexes, such as that for complex 12,
A common approach for evaluating the oxidation state of metals have well-resolved pre-edge features: a weak feature at ~8333
in diphosphine complexes involves using the magnitude of P,P- eV and a more intense feature at ~8336 eV.
scalar coupling constants (2JP,P). In nickel chemistry, it is generally The weaker feature results from the electric-quadrupole allowed
observed that small 2JP,P (i.e. 2-30 Hz) correspond to nickel(II) Ni 1s → 3d transition, whereas the more intense feature in such
complexes whereas larger 2JP,P (i.e. 45-80 Hz) correspond to complexes has previously been ascribed to a dipole-allowed Ni
nickel(0) complexes.[10,11,22,77–84] However, exceptions to this trend 1s → 4p transition .[98,99] By contrast, Ni π-complexes (such as 1,
have been reported by ourselves[10] and others.[84,85] Moreover, 2, 4, and 7) have a markedly different edge profile and only one
this approach is limited to asymmetric species as 2JP,P cannot be clearly resolvable intense pre-edge feature ranging from 8333-
observed in complexes such as 4 and 7 due to symmetry. 8336 eV. Similar spectroscopic behaviour has previously been
Others have noted similar results with infrared (IR) spectroscopy observed in copper(I)-derived π-complexes.[87] As expected, the
of nickel carbonyl complexes[73] as we have found from XRD and energy of the Ni 1s → 4p feature correlates with the oxidation
NMR, i.e. that electron donation from the metal to the ligand does state of the metal centre. The weaker pre-edge feature is not
reduce the C=O bond, but to an extent that is ambiguously directly resolvable in most complexes, although in complexes 1 &
between a single and double bond. Importantly, none of these 2 a weak low-energy shoulder is observed in the 2nd derivative of
traditionally-used methods directly provides information about the the spectra (see SI 04).
Ni Kβ X-ray emission spectroscopy 1𝑠𝑠) shifts in the opposite direction, such that the energy difference
between the two features (∆𝐸𝐸𝑑𝑑𝑑𝑑 ) increases with increasing
To further explore the charge distribution, Ni Kβ emission data was
oxidation at the metal centre. The weak low-energy 3d feature
obtained for select complexes. The intensity-weighted average
should eventually be unresolvable from the higher intensity 4p
energy of the Kβ line for each of the analysed complexes suggests
feature, as observed in the experimental data.
a similar Zeff (see SI 35). This trend is similar to that obtained from
the energies of the Ni 1s → 4p feature in the Ni K-edge XAS.
Indeed, the emission energies for the π-complexes 1-7 are quite
similar (± 1 eV) and well separated from the observed energy of
authentic Ni(II) complex 12. The similarity in the Kβ emission data
for all of the π-complexes seems to support a common formal
oxidation state for 1-7.

Computational Studies - Density Functional Theory

DFT calculations were performed on each of the species in Chart
1 using simplified diphosphine ligands (See SI 01 for details).
Molecular structures derived from B3LYP/def2-TZVP calculations
of the dtbpe complexes, as well as those using a simplified
diphosphine ligand (dmpe = 1,2-bis(dimethylphosphino) ethane),A
yield good agreement with solid-state molecular structures of 1-
12.[10,11,22,57,100,101] The effect of decreasing the steric bulk and Figure 3. Simplified MO diagram depicting differences between weaker (left,
electron donation in the supporting diphosphine ligand does not e.g. ethylene in 4) and stronger π-acidic ligands (right, e.g. cyclohexanone in
∗
6). Greater π-acidity leads to a much lower 𝜋𝜋𝑖𝑖𝑖𝑖 and thus greater π-backbonding.
affect the general structural trends and conclusions, which are
Decreased electron density at the metal centre (i.e. increased 𝑍𝑍𝑒𝑒𝑒𝑒𝑒𝑒 ) also lowers
consistent with those observed in the experimental data. the energy of the Ni 1s orbital. These two effects lead to a simultaneous
Furthermore, the spectroscopic features observed in the Ni K- increase in energy of the Ni 1s → 4p transition (red arrow) and decrease in
edge XAS data are well reproduced using TD-DFT analysis. energy of the Ni 1s→ π* (blue arrow) and therefore an increase in the splitting
of the two acceptor orbitals (∆𝐸𝐸𝑑𝑑𝑑𝑑 ). Quantitative results are given in supporting
Although qualitative results were consistent across a broad range
information as SI 41.
of functionals, results from B3LYP provided the best agreement
with experimental pre-edge features. Basis set effects were
observed to be minimal beyond TZVP. The strong agreement with
The nature of the two pre-edge final states is consistent across
experimental data suggests that our DFT results should provide a
the series of π-complexes. The intense feature results from a
reasonable description of bonding in these species.
transition to the non-bonding Ni 4pz orbital, whose energy directly
reflects 𝑍𝑍𝑒𝑒𝑒𝑒𝑒𝑒 at the metal centre. The weaker feature corresponds
to a transition to the formally ligand-based antibonding π* orbital,
which gains electric quadrupole character by mixing with the Ni
3dx2-y2 orbital through π-backbonding. In principle, the intensity of
this features should therefore reflect the degree of M-L π-
backbonding. However, the intensity of this pre-edge feature also
depends on the degree of Ni d-p mixing, which varies across the
series. This contribution, in addition to the difficulty of resolving
weak pre-edge shoulder, makes it challenging to quantify the
degree of backbonding from the experimental data.
To further explore the electron distribution in the ground state of
these π-systems, we applied charge decomposition (CDA),[102,103]
natural bond orbital (NBO),[104] and quantum theory of atoms-in-
molecules (QTAIM)[105] analyses. Together, these provide a
comprehensive view of the electronic properties of these systems.
In all cases, the predominant interactions between the metal ion
Figure 2. Calculated Ni K-edge XANES TD-DFT results for pre-edge region of and the π-ligand can be well described using the basic
the spectrum. Each complex is represented by a blue circle (Ni 4pz←1s) and a
red circle (Ni 3dz2-y2←1s). The area of the of each circle is proportional to the interactions defined within the DCD bonding model.
calculated oscillator strength (fosc) for each transition. All calculated TD-DFT Most notably, π-backbonding is the dominant contribution to
energies at the Ni K-edge were shifted by -98.55 eV. [Ni] = Ni(dmpe). bonding in these systems. The electron rich metal centre does not
accept significant electron density through σ donation from the πb
ligand orbital with only minimal charge donation into the higher
The Ni K-edge pre-edge features are extremely sensitive to lying empty Ni 4s/p orbitals. The π-backbonding interaction
electron distribution (Figure 2). The more intense, higher energy involves overlap between the Ni 3dx2-y2 and the in-plane ligand 𝜋𝜋 ∗
(𝑁𝑁𝑁𝑁 4𝑝𝑝𝑧𝑧 ← 1𝑠𝑠) transition increases with greater oxidation at the ∗
(𝜋𝜋𝑖𝑖𝑖𝑖 ). As expected, the overall degree of charge transfer
metal centre and reproduces the trend observed in the correlates directly with the relative energies of these contributing
experimental data. The weaker, low energy transition (𝑁𝑁𝑁𝑁 3𝑑𝑑 ← fragment orbitals (see SI 41). Given that the Ni(dtbpe) fragment is
identical in all cases, differences within the series result primarily of the DFT-derived electron density (∇2 (𝜌𝜌𝐷𝐷𝐷𝐷𝐷𝐷 )) for 4’ and 1’. In the
∗
from changes in the energy of the ligand 𝜋𝜋𝑖𝑖𝑖𝑖 orbital. As olefinic π-complex, the electron density within the Ni-C-C trigonal
summarized in Figure 3, poorer π-acids such as olefins have a core reveals two Ni-C bond critical points (bcp) and one ring
∗ and thus should exhibit a small ∆𝐸𝐸𝑑𝑑𝑑𝑑 , whereas
higher energy 𝐸𝐸𝜋𝜋𝑖𝑖𝑖𝑖 critical point (rcp) that connects all three atoms. The rcp correlates
∗ is lower in energy for stronger π-acids (such as carbonyls),
𝐸𝐸𝜋𝜋𝑖𝑖𝑖𝑖 with a σ donor interaction due to πCC donation in the Ni 3dxy orbital
increasing backbonding and a larger ∆𝐸𝐸𝑑𝑑𝑑𝑑 .This interaction leads and the two Ni-C bcp’s correspond to π-backbonding from the Ni
∗
to a surprisingly large barrier for ligand rotation, even for those 3dx2-y2 and the ligand 𝜋𝜋𝑖𝑖𝑖𝑖 . By contrast, ∇2 (𝜌𝜌𝐷𝐷𝐷𝐷𝐷𝐷 ) for 1’ is highly
where backbonding is least important: barriers of ~100 kJ/mol are asymmetric with two bcp’s (Ni-O and Ni-C) but no discernible rcp
obtained for both symmetric (4) and asymmetric (3) π-ligands. in this case.
Table 1. Wiberg bond indices for Ni-C, Ni-X (X=O or most electron-rich C), and
QTAIM ∇2(ρDFT) for optimized complexes at B3LYP/def2-TZVP level of theory.
Discussion
Wiberg Indices 𝛁𝛁 𝟐𝟐 (𝝆𝝆𝑫𝑫𝑫𝑫𝑫𝑫 )

Ni-C Ni-X bcpNiC bcpNiX rcpNiCO Our studies of a series of nickel π-complexes reveal interesting
7’ 0.362 0.357 0.217 0.224 0.309 electronic structure features that can be rationalized within the
context of the DCD bonding model. The spectroscopic
C=C 5’ 0.495 0.403 0.213 0.261 0.353
characteristics of these species are highly sensitive to the nature
4’ 0.486 0.487 0.235 0.238 0.359
of bonding to the π-ligand, more specifically the properties of the
3’ 0.516 0.487 0.524 0.248 - species are intimately linked to the degree of π-backbonding from
C=O 1’ 0.556 0.468 0.550 0.219 - the electron-rich metal centre. Taken together, our studies allow
for a more concrete evaluation of the factors that control this
2’ 0.552 0.502 0.537 0.243 -
bonding and their implications.
6’ 0.523 0.494 0.536 0.232 -
Although the above analysis is valid for all species investigated,
there is one additional factor that contributes to the nature of the
bonding in these systems. The significant electronegativity
difference between carbon and oxygen in the carbonyl π-ligands
leads to asymmetry in the orbitals involved in bonding. Indeed,
this bonding difference between olefins and carbonyls was
identified by Eisenstein and Hoffmann nearly four decades ago.[73]
The nature of bonding in these asymmetric systems is therefore
more complex and deviates somewhat from the simple DCD
model as σ donation becomes more localized from the terminal
oxygen atom and π-backbonding localizes onto the electron
deficient carbonyl carbon atom. This localization is also consistent
with π-backbonding (to C) being stronger than σ-donation (from
O), as observed from bond strength parameters in Table 1.

Figure 5. Correlation between Ni 1s→4p transition energies and 2JPP NMR

coupling constants. Data points in black circles are from TDDFT calculations
whereas those in red circles are from experimental Ni K-edge XAS data. All TD-
DFT calculated transition energies were linearly shifted by -98.55 eV. The
dashed line represents a linear correlation fit (𝑅𝑅2 = 0.87); see SI 3.

Oxidation states of nickel diphosphines are often evaluated via

the magnitude of 2𝐽𝐽𝑃𝑃,𝑃𝑃 in unsymmetrical complexes: a small
coupling constant (2 𝐻𝐻𝐻𝐻 < 2𝐽𝐽𝑃𝑃,𝑃𝑃 < 30 𝐻𝐻𝐻𝐻) correlates with Ni(II)
Figure 4. QTAIM topological analysis for complexes 4’ (left) and 1’ (right).
Contour maps of ∇2 (𝜌𝜌𝐷𝐷𝐷𝐷𝐷𝐷 ) in the NiCX plane (X=C, O). Dotted contours refer to
complexes, whereas large coupling constants (45 𝐻𝐻𝐻𝐻 < 2𝐽𝐽𝑃𝑃,𝑃𝑃 <
positive values of ∇2 (𝜌𝜌𝐷𝐷𝐷𝐷𝐷𝐷 ) and solid lines to negative values of ∇2 (𝜌𝜌𝐷𝐷𝐷𝐷𝐷𝐷 ). Bond 80 𝐻𝐻𝐻𝐻) are associated with Ni(0) species. Differences reflect the
critical points are shown in blue and ring critical points are shown in red. A electron density at the metal centre, which bridges the two 31P
simplified representation of these bonding interactions is shown on the bottom atoms.[106] There is little evidence regarding potential Ni(I) species
left of for each of the complexes.
given the challenges associated with obtaining such information
in paramagnetic species.[13,81] XAS offers the advantage of
providing an independent experimental probe requiring neither
The effect of π-ligand asymmetry is also clearly observed in the
inequivalent phosphorus ligands nor diamagnetism. As noted
QTAIM analysis: Figure 4 shows a comparison of the Laplacian
previously, the XAS data of formally square planar complexes
yield distinctive pre-edge features that track with oxidation of the
metal centre. The NMR spectroscopy coupling constants and
XAS pre-edge energies correlate extremely well (Figure 5),
providing good support that 2𝐽𝐽𝑃𝑃,𝑃𝑃 (where available) are useful in
defining electron density at the metal centre.
The fact that the two pre-edge features in the Ni K-edge XAS data
respond so differently to changes in the electronic structure
implies that they are sensitive to different aspects of the electronic
structure of the metal centre. The Ni 4pz orbital is out-of-plane
from the most important ligand field interactions in pseudo square
planar geometries and thus reports directly on Zeff of the metal
centre. In contrast, the weak pre-edge feature is a predominantly
in-plane ligand-based final state with some metal 3𝑑𝑑𝑥𝑥 2−𝑦𝑦2
character. The two features therefore behave very differently with
the former reporting on Zeff of the metal centre and the latter on
differences in the ligand field.
The DCD model is a simple yet powerful approach for explaining
the behaviour of π-complexes in transition metal chemistry. Its
limitations have recently been explored in copper dioxygen and
related systems by invoking the important contributions of static
correlation, specifically by allowing for multi-determinant
solutions.[87] Since the electron density of these systems are well
described from DFT calculations, we approached this same issue
by applying natural resonance theory[107–109] (NRT) to expose
different contributions to the overall electronic description (Table
2).[8] In all cases, the Ni(II) metallacycle contributes little to the
overall electronic structure. The Ni(0) π-adduct and Ni(I)
intermediate resonance structures account for >80% of the Figure 6. (A) DFT-calculated charge donation from CDA analysis for ethylene
complex with dmpe ligand (4’) for ground state geometry (𝜙𝜙 = 0°) and after
electronic structure in all cases. Indeed, we find that the Ni(0) π-
tetragonal distortion (rotation of NiCC plane relative to NiPP plane, 𝜙𝜙 = 90°).
adduct is the largest contributor for all the structures examined, Backbonding decreases substantially due to large drop in phosphine σ donation
although Ni(I) contributions are non-negligible. into the Ni 3dx2-y2. (B) Qualitative molecular orbital representation of the 3c4e
∗
bonding that connects 𝑃𝑃𝜎𝜎− with 𝜋𝜋𝑖𝑖𝑖𝑖 via the Ni 3𝑑𝑑𝑥𝑥 2−𝑦𝑦2 orbital.
Table 2. Summary of NRT analyses for complexes with either olefin or carbonyl
π-ligands.

Ni(0) Ni(I) Ni(II) Charge donation from the P2Ni fragment to the π-ligand
7’ 67% 33% 0% decreases substantially when the diphosphine ligand is
perpendicular to the Ni 3𝑑𝑑𝑥𝑥 2−𝑦𝑦2 donor orbital (Figure 6A). This
C=C 5’ 56% 37% 7%
suggests that the metal centre in these complexes mediates
4’ 51% 36% 10%
charge donation from the electron rich diphosphine to the
3’ 57% 38% 5% ethylene π*. The orbital contributions that allow for π-
C=O 1’ 48% 47% 4% backbonding are reminiscent of a classical 3c-4e bond; in this
situation the three contributing orbitals are the antisymmetric
2’ 46% 43% 11%
combination of the phosphine σ-donor orbitals (𝑃𝑃− 𝜎𝜎 , 2 valence
∗
electrons), the Ni 3𝑑𝑑𝑥𝑥 2−𝑦𝑦2 (2 valence electrons), and the ligand 𝜋𝜋𝑖𝑖𝑖𝑖
The high barrier to rotation for π-ligands in these complexes
orbital (Figure 6B). In this geometry, the two sets of ligands
implies a surprisingly strong preference for a planar geometry
generate a cooperative “push-pull” system mediated by the metal
even though a closed-shell Ni(0) π-adduct should not behave in
centre in a manner similar to that which has been observed in
this way. Even more surprisingly, the barrier to rotation does not
cytochrome P450s.[110] This 𝜎𝜎 → 𝜋𝜋 cooperative interaction
correlate strongly with the degree of Ni(I) character from our NRT
mediation by the metal centre is only possible in specific
analysis. This effect points to the importance of the trans-
geometries and is essentially identical to mixed σ/π interactions
diphosphine ligand in enabling and supporting π-backbonding. In
observed in trigonal systems.(ref – π contributions in tbp
principle, π-backbonding in a d10 Ni(0) occurs in any geometry of
complexes). This would preclude similar effects with ancillary
the π-ligand because of availability of filled Ni 3dxz,yz orbitals that
ligands that enforce a linear geometry in Ni(0) complexes, such
could also support backbonding. However, the C=C bond
as N-heterocyclic carbenes (NHCs). It is noteworthy that Itami[111]
distance (rCC) decreases significantly upon ligand rotation (from
and Liu[112] have identified striking reactivity differences in the
147 to 139 pm), indicating that backbonding is not well supported
catalytic activity of monodentate phosphines versus their
in alternate geometries. The electronic changes that occur upon
bidentate analogues, with additional computational studies
rotation of the π-ligand are an indicator of the importance of the
performed by Houk.[112]
diphosphine ligand.
Although all of the olefin complexes investigated are best Experimental Section
described as Ni(0) complexes, highly electron poor carbonyl
complexes have significantly more oxidized metal centres. General Considerations
Complexes 1 and 2 represent intriguing examples of intermediate
cases that are shifting towards a “Ni(I)”-type description. The CF3 Unless stated otherwise, all reactions were performed in a glovebox or on
substituent at the carbonyl carbon increases their π-acidity a Schlenk line under an atmosphere of pure N2 using standard Schlenk
substantially. However, the ester ligand in 1 is less π-acidic than techniques. Anhydrous pentanes, toluene, diethyl ether, and
the equivalent thioester in 2 due to better delocalization of the tetrahydrofuran were purchased from Aldrich, sparged with N2, and dried
further by passage through towers containing activated alumina and
ligand π-system, which simultaneously decreases overlap of the
molecular sieves. C6H6 and C6D6 were purchased from Aldrich and dried
π* with the Ni 3dx2-y2 and increases its energy. In 2, poor π-overlap over sodium/benzophenone before being distilled and degassed by three
between the larger atomic orbitals on sulphur and the carbonyl π* freeze-pump-thaw cycles. CD2Cl2 was purchased from Aldrich and dried
allows for significant backbonding in this case. These differences over CaH2 before being distilled and degassed by three freeze-pump-thaw
should lead to concomitant differences in reactivity. cycles. Ni(COD)2 (13) was purchased from Strem and used as received.
Indeed, 1 and 2 display fundamentally different reactivity. In Compounds 1-2,[11] 3,[57] 4,[100] 5-6,[10] 7,[115] 8-10,[11] 11,[10] 12,[113] 14,[116]
refluxing benzene, 2 slowly thermolyses over two days, resulting and 15[100] were prepared according to literature procedures. All other
in complex 15, free ligand, and thioester, as determined by chemicals were purchased from commercial suppliers and used as
31 received.
P{1H} and 1H NMR spectroscopy. In contrast, complex 1 is
stable for up to a week under the same conditions with no sign of
X-ray Absorption Spectroscopy
decomposition. Complex 1 does not react with MeLi, even upon
prolonged reflux in benzene, but under the same conditions,
All the XAS samples were analysed as solids under anaerobic conditions
complex 2 reacts with MeLi to form trace amounts of EtSSEt. In and diluted in boron nitride (20-50% by weight). XAS Ni K-edges were
addition, complex 2 is susceptible to cross-coupling with acquired at the SSRL beamline 7-3, which is equipped with a Si(220) ϕ =
phenylboronic acid, forming PhSEt in moderate (35%) yield.[11] No 90° double crystal monochromator, a 9 keV cutoff mirror, and a He cryostat
such cross-coupling reactivity was observed with complex 1. (at 20 K). Data were collected using a Canberra 30-element Ge solid-state
Lastly, complexes 1 and 2 react differently with MeI. Upon detector with a 3mm Co filter. Data averaging and energy calibration were
refluxing in benzene for 12 hours, 1 forms (dtbpe)Ni(Me)(I),[113] performed using SixPack[117] and the AUTOBK algorithm available in the
which was verified by 31P NMR spectroscopy, and liberates the Athena software package[118] was employed for data reduction and
normalization. Independent fitting was also performed using
free ester, whereas complex 2 does not react with MeI at all under
BlueprintXAS.[119,120]
the same conditions.[B] Such behaviour is consistent with reduced
π-donation of the ester in 1, allowing for its displacement by MeI.
X-ray emission spectroscopy

Samples were prepared for XES by pressing finely ground powders into 1
Conclusions mm Al spacers, and sealing with 40 μm Kapton tape. Data were obtained
at the Cornell High Energy Synchrotron Radiation Source (CHESS) at the
Our spectroscopic and computational studies on a series of Ni π- C-line end station. Energy selection was performed with upstream
complexes shed light on an intriguing effect whereby σ-donor multilayers, providing ∼50 eV band pass. A Rh-coated harmonic rejection
ancillary ligands are instrumental in stabilizing a square-planar mirror was also utilized. Kβ X-ray emission spectra were measured using
a spherical analyzer (using the 620 reflection of three Ge 310 analyzer
geometry in formally d10 nickel(0) complexes. In the case of
crystals) in combination with a silicon drift detector aligned in Rowland
olefinic complexes, the evidence shows that these are best geometry. Data were normalized with respect to the incident flux in an
described as Ni(0) π-adducts with strong π-backbonding coupled upstream N2-filled ionization chamber. Data were collected at ∼20 K in a
to in-plane σ donation from the supporting electron rich Displex cryostat to minimize photoreduction.
diphosphine bidentate ligand. The formation of this 3c-4e
interaction generates ligand σ-to-π (Lσ→Lπ) charge transfer. This Computational methods
ligand-induced (push-pull) π-backbonding is responsible for the
large observed rotational barrier about the π-ligand even with Initial geometries for all molecules were obtained from crystallographic
relatively poor π-acidic ligands such as ethylene. This unique coordinates (where available) or constructed from standard models.
electronic structure can play an important role in the reactivity of Geometry optimizations and numerical frequency calculations were
such species.[114] The situation is more complex in situations with performed using version 3.0.3 of the ORCA computational chemistry
package. Molecular geometries were optimized using the B3LYP
highly electron poor π-systems, where metal-centred
functional in combination with the Ahlrichs triple-ζ basis set with valence
backbonding increases and leads to Ni(I) character becomes polarization (def2-TZVP) for all atoms. Computational efficiency was
significantly more important. The implication here for catalyst improved by applying the RI approximation (RIJCOSX) for the hybrid
design is that the orbital energies of the unsaturated substrate can functional. All calculations were performed with integration grid 4.
be matched not only with the metal centre but with the Reported thermochemical energies are given in kJ/mol and correspond to
diphosphine ligand. We anticipate that the tuning of both electron- Gibbs free energies (ΔG0) with zero-point vibrational energy corrections
donating diphosphine and the electron-accepting π-ligand will (ZPVE). NBO results were obtained using Gaussian 09; AIM and CDA
allow for an additional handle in the design of subsequent calculation were performed in Multiwfn software from NBO outputs. All
calculations were run on either the Abacus (UBC Chemistry) or GREX
catalysts, and that more detailed studies of ancillary ligand effects
(Westgrid) computing clusters.
in such systems are warranted.
Acknowledgements [19] B. A. Vara, X. Li, S. Berritt, C. R. Walters, E. J. Petersson, G. A.
Molander, Chem. Sci. 2018, 9, 336–344.
[20] B. P. Woods, M. Orlandi, C.-Y. Huang, M. S. Sigman, A. G. Doyle, J. Am.
This work was supported by the Natural Sciences and
Chem. Soc. 2017, 139, 5688–5691.
Engineering Research Council of Canada (NSERC) via Discovery [21] C. Heinz, J. P. Lutz, E. M. Simmons, M. M. Miller, W. R. Ewing, A. G.
grants and Research Tools and Instrumentation grants to J.A.L. Doyle, J. Am. Chem. Soc. 2018, 140, 2292–2300.
and P.K. Scholarship support was provided to A.N.D. by NSERC [22] M. L. Lejkowski, R. Lindner, T. Kageyama, G. É. Bódizs, P. N. Plessow,
(CGS-D3 + MSFSS), the University of British Columbia (Laird + I. B. Müller, A. Schäfer, F. Rominger, P. Hofmann, C. Futter, et al., Chem.
4YF), and the Izaak Walton Killam Foundation during his graduate Eur. J. 2012, 18, 14017–14025.
studies. W.H. is grateful for support in the form of a MITACS [23] N. Saito, Z. Sun, Y. Sato, Chem. Asian J. 2015, 10, 1170–1176.
[24] S. C. E. Stieber, N. Huguet, T. Kageyama, I. Jevtovikj, P. Ariyananda, A.
Globalink doctoral fellowship. W.C. is grateful for support from the
Gordillo, S. A. Schunk, F. Rominger, P. Hofmann, M. Limbach, Chem.
NSERC CREATE SusSyn program.
Commun. 2015, 51, 10907–10909.
This work is partially based upon research performed at the [25] M. Al-Ghamdi, S. V. C. Vummaleti, L. Falivene, F. A. Pasha, D. J.
Stanford Synchrotron Radiation Lightsource (SSRL) at the SLAC Beetstra, L. Cavallo, Organometallics 2017, 36, 1107–1112.
National Accelerator Laboratory. Use of SSRL is supported by the [26] Z. R. Greenburg, D. Jin, P. G. Williard, W. H. Bernskoetter, Dalton Trans.
U.S. Department of Energy, Office of Science, Office of Basic 2014, 43, 15990–15996.
Energy Sciences under Contract No. DE-AC02-76SF00515. This [27] I. Knopf, D. Tofan, D. Beetstra, A. Al-Nezari, K. Al-Bahily, C. C.
work is partially based upon research conducted at the Cornell Cummins, Chem. Sci. 2017, 8, 1463–1468.
[28] C. Hendriksen, E. A. Pidko, G. Yang, B. Schäffner, D. Vogt, Chem. Eur.
High Energy Synchrotron Source (CHESS) which is supported by
J. 2014, 20, 12037–12040.
the National Science Foundation under award DMR-1332208.
[29] N. Huguet, I. Jevtovikj, A. Gordillo, M. L. Lejkowski, R. Lindner, M. Bru,
This research was enabled in part by support provided by A. Y. Khalimon, F. Rominger, S. A. Schunk, P. Hofmann, et al., Chem.
WestGrid (www.westgrid.ca) and Compute Canada Calcul Eur. J. 2014, 20, 16858–16862.
Canada (www.computecanada.ca). [30] D. Jin, P. G. Williard, N. Hazari, W. H. Bernskoetter, Chem. Eur. J. 2014,
20, 3205–3211.
[31] S. Manzini, N. Huguet, O. Trapp, T. Schaub, Eur. J. Org. Chem. 2015,
[A] Computed structures using the dmpe ligand are labeled using a ‘ (prime) 2015, 7122–7130.
in the numbering. For example, 4’ is simply complex 4 from Chart 1 that [32] W. Guo, C. Michel, R. Schwiedernoch, R. Wischert, X. Xu, P. Sautet,
has been calculated using the simplified dmpe ligand rather than the full Organometallics 2014, 33, 6369–6380.
dtbpe ligand. [33] S. Ogoshi, Y. Hoshimoto, M. Ohashi, Chem. Commun. 2010, 46, 3354.
[B] Both 1 and 2 react with HCl to make (dtbpe)NiCl2 (12) and PhLi to form [34] Y. Hoshimoto, M. Ohashi, S. Ogoshi, J. Am. Chem. Soc. 2011, 133,
moderate but irreproducible amounts of biphenyl. 4668–4671.
[1] O. R. Luca, R. H. Crabtree, Chem. Soc. Rev. 2013, 42, 1440–1459. [35] Y. Hoshimoto, M. Ohashi, S. Ogoshi, Acc. Chem. Res. 2015, 48, 1746–
[2] V. Lyaskovskyy, B. de Bruin, ACS Catal. 2012, 2, 270–279. 1755.
[3] P. J. Chirik, K. Wieghardt, Science 2010, 327, 794–795. [36] R. Kumar, Y. Hoshimoto, H. Yabuki, M. Ohashi, S. Ogoshi, J. Am. Chem.
[4] M. W. Bouwkamp, A. C. Bowman, E. Lobkovsky, P. J. Chirik, J. Am. Soc. 2015, 137, 11838–11845.
Chem. Soc. 2006, 128, 13340–13341. [37] Y. Hoshimoto, H. Yabuki, R. Kumar, H. Suzuki, M. Ohashi, S. Ogoshi, J.
[5] H. Nishiyama, H. Ikeda, T. Saito, B. Kriegel, H. Tsurugi, J. Arnold, K. Am. Chem. Soc. 2014, 136, 16752–16755.
Mashima, J. Am. Chem. Soc. 2017, 139, 6494–6505. [38] Y. Bernhard, B. Thomson, V. Ferey, M. Sauthier, Angew. Chem. 2017,
[6] K. J. Blackmore, J. W. Ziller, A. F. Heyduk, Inorg. Chem. 2005, 44, 5559– 56, 7460–7464.
5561. [39] Y. Kita, R. D. Kavthe, H. Oda, K. Mashima, Angew. Chem. 2016, 55,
[7] B. de Bruin, M. J. Boerakker, J. J. J. M. Donners, B. E. C. Christiaans, P. 1098–1101.
P. J. Schlebos, R. de Gelder, J. M. M. Smits, A. L. Spek, A. W. Gal, [40] Y. Kita, H. Sakaguchi, Y. Hoshimoto, D. Nakauchi, Y. Nakahara, J.-F.
Angew. Chem. 1997, 36, 2064–2067. Carpentier, S. Ogoshi, K. Mashima, Chem. Eur. J. 2015, 21, 14571–
[8] A. N. Desnoyer, S. Behyan, B. O. Patrick, A. Dauth, J. A. Love, P. 14578.
Kennepohl, Inorg. Chem. 2016, 55, 13–15. [41] M. S. Azizi, Y. Edder, A. Karim, M. Sauthier, Eur. J. Org. Chem. 2016,
[9] P. H. M. Budzelaar, A. N. J. Blok, Eur. J. Inorg. Chem. 2004, 2004, 2385– 2016, 3796–3803.
2391. [42] D. K. Nielsen, A. G. Doyle, Angew. Chem. 2011, 50, 6056–6059.
[10] A. N. Desnoyer, E. G. Bowes, B. O. Patrick, J. A. Love, J. Am. Chem. [43] K. Muto, J. Yamaguchi, D. G. Musaev, K. Itami, Nat. Commun. 2015, 6,
Soc. 2015, 137, 12748–12751. 7508.
[11] A. N. Desnoyer, F. W. Friese, W. Chiu, M. W. Drover, B. O. Patrick, J. A. [44] M. L. Smith, A. K. Leone, P. M. Zimmerman, A. J. McNeil, ACS Macro
Love, Chem. Eur. J. 2016, 22, 4070–4077. Lett. 2016, 5, 1411–1415.
[12] N. A. LaBerge, J. A. Love, Eur. J. Org. Chem. 2015, 2015, 5546–5553. [45] A. K. Leone, A. J. McNeil, Acc. Chem. Res. 2016, 49, 2822–2831.
[13] D. D. Beattie, E. G. Bowes, M. W. Drover, J. A. Love, L. L. Schafer, [46] B. C. Achord, J. W. Rawlins, Macromolecules 2009, 42, 8634–8639.
Angew. Chem. 2016, 55, 13290–13295. [47] R. Miyakoshi, A. Yokoyama, T. Yokozawa, J. Am. Chem. Soc. 2005, 127,
[14] A. N. Desnoyer, J. Geng, M. W. Drover, B. O. Patrick, J. A. Love, Chem. 17542–17547.
Eur. J. 2017, 23, 11509–11512. [48] M. C. Iovu, E. E. Sheina, R. R. Gil, R. D. McCullough, Macromolecules
[15] D. R. Hartline, M. Zeller, C. Uyeda, J. Am. Chem. Soc. 2017, 139, 13672– 2005, 38, 8649–8656.
13675. [49] P. Willot, G. Koeckelberghs, Macromolecules 2014, 47, 8548–8555.
[16] D. J. Mindiola, Angew. Chem. 2009, 48, 6198–6200. [50] S. R. Lee, J. W. G. Bloom, S. E. Wheeler, A. J. McNeil, Dalton Trans.
[17] P. Zimmermann, C. Limberg, J. Am. Chem. Soc. 2017, 139, 4233–4242. 2013, 42, 4218.
[18] E. B. Corcoran, M. T. Pirnot, S. Lin, S. D. Dreher, D. A. DiRocco, I. W. [51] Z. J. Bryan, A. J. McNeil, Chem. Sci. 2013, 4, 1620.
Davies, S. L. Buchwald, D. W. C. MacMillan, Science 2016, 353, 279– [52] S. R. Lee, Z. J. Bryan, A. M. Wagner, A. J. McNeil, Chem. Sci. 2012, 3,
283. 1562.
[53] E. L. Lanni, J. R. Locke, C. M. Gleave, A. J. McNeil, Macromolecules [89] S. N. MacMillan, K. M. Lancaster, ACS Catal. 2017, 7, 1776–1791.
2011, 44, 5136–5145. [90] J. L. DuBois, P. Mukherjee, T. D. P. Stack, B. Hedman, E. I. Solomon, K.
[54] H. Komber, V. Senkovskyy, R. Tkachov, K. Johnson, A. Kiriy, W. T. S. O. Hodgson, J. Am. Chem. Soc. 2000, 122, 5775–5787.
Huck, M. Sommer, Macromolecules 2011, 44, 9164–9172. [91] J. Kowalska, S. DeBeer, Biochim. Biophys. Acta - Mol. Cell Res. 2015,
[55] R. Tkachov, V. Senkovskyy, H. Komber, J.-U. Sommer, A. Kiriy, J. Am. 1853, 1406–1415.
Chem. Soc. 2010, 132, 7803–7810. [92] T. Yamamoto, X-Ray Spectrom. 2008, 37, 572–584.
[56] E. L. Lanni, A. J. McNeil, J. Am. Chem. Soc. 2009, 131, 16573–16579. [93] A. Kuzmin, N. Mironova, J. Purans, A. Rodionov, J. Phys. Condens.
[57] A. N. Desnoyer, W. Chiu, C. Cheung, B. O. Patrick, J. A. Love, Chem. Matter 1995, 7, 9357–9368.
Commun. 2017, 53, 12442–12445. [94] A. Corrias, G. Mountjoy, G. Piccaluga, S. Solinas, J. Phys. Chem. B
[58] G. Erker, U. Dorf, P. Czisch, J. L. Peterson, Organometallics 1986, 5, 1999, 103, 10081–10086.
668–676. [95] A. Rougier, C. Delmas, A. V. Chadwick, Solid State Commun. 1995, 94,
[59] G. Erker, U. Dorf, J. L. Atwood, W. E. Hunter, J. Am. Chem. Soc. 1986, 123–127.
108, 2251–2257. [96] I. J. Pickering, G. N. George, J. T. Lewandowski, A. J. Jacobson, J. Am.
[60] F. R. Askham, K. M. Carroll, S. J. Alexander, A. L. Rheingold, B. S. Chem. Soc. 1993, 115, 4137–4144.
Haggerty, Organometallics 1993, 12, 4810–4815. [97] A. N. Mansour, C. A. Melendres, J. Phys. Chem. A 1998, 102, 65–81.
[61] J. E. Hill, P. E. Fanwick, I. P. Rothwell, Organometallics 1992, 11, 1771– [98] C. Gougoussis, M. Calandra, A. Seitsonen, C. Brouder, A. Shukla, F.
1773. Mauri, Phys. Rev. B - Condens. Matter Mater. Phys. 2009, 79, 045118.
[62] G. Fachinetti, C. Biran, C. Floriani, A. Chiesi-Villa, C. Guastini, J. Am. [99] J. Cho, R. Sarangi, J. Annaraj, S. Y. Kim, M. Kubo, T. Ogura, E. I.
Chem. Soc. 1978, 100, 1921–1922. Solomon, W. Nam, Nat. Chem. 2009, 1, 568–572.
[63] L. Li, K. E. Kristian, A. Han, J. R. Norton, W. Sattler, Organometallics [100] K.-R. Pörschke, C. Pluta, B. Proft, F. Lutz, C. Krüger, Zeitschrift für
2012, 31, 8218–8224. Naturforsch. B 1993, 48, 608–626.
[64] P. T. Wolczanski, Organometallics 2017, 36, 622–631. [101] M. Schultz, P.-N. Plessow, F. Rominger, L. Weigel, Acta Crystallogr.
[65] G. Parkin, J. Chem. Educ. 2006, 83, 791. Sect. C Cryst. Struct. Commun. 2013, 69, 1437–1447.
[66] M. L. H. Green, G. Parkin, J. Chem. Educ. 2014, 91, 807–816. [102] M. Joost, L. Estévez, S. Mallet-Ladeira, K. Miqueu, A. Amgoune, D.
[67] L. Yang, D. R. Powell, R. P. Houser, Dalton Trans. 2007, 955–964. Bourissou, Angew. Chem. 2014, 53, 14512–14516.
[68] M. Benedetti, C. R. Barone, D. Antonucci, V. M. Vecchio, A. Ienco, L. [103] S. Dapprich, G. Frenking, J. Phys. Chem. 1995, 99, 9352–9362.
Maresca, G. Natile, F. P. Fanizzi, Dalton Trans. 2012, 41, 3014. [104] A. E. Reed, L. a Curtiss, F. Weinhold, Chem. Rev. 1988, 88, 899–926.
[69] “Wadepohl CCDC private communication,” n.d. [105] L. J. Farrugia, C. Evans, D. Lentz, M. Roemer, J. Am. Chem. Soc. 2009,
[70] B. de Bruin, M. J. Boerakker, J. A. W. Verhagen, R. de Gelder, J. M. M. 131, 1251–1268.
Smits, A. W. Gal, Chem. Eur. J. 2000, 6, 298–312. [106] R. D. Bertrand, F. B. Ogilvie, J. G. Verkade, J. Am. Chem. Soc. 1970,
[71] L. E. Doyle, W. E. Piers, J. Borau-Garcia, M. J. Sgro, D. M. Spasyuk, 92, 1908–1915.
Chem. Sci. 2016, 7, 921–931. [107] E. D. Glendening, F. Weinhold, J. Comput. Chem. 1998, 19, 593–609.
[72] L. E. Doyle, W. E. Piers, J. Borau-Garcia, J. Am. Chem. Soc. 2015, 137, [108] E. D. Glendening, F. Weinhold, J. Comput. Chem. 1998, 19, 610–627.
2187–2190. [109] E. D. Glendening, J. K. Badenhoop, F. Weinhold, J. Comput. Chem.
[73] J. Kaiser, J. Sieler, D. Walther, E. Dinjus, L. Golic, Acta Crystallogr. Sect. 1998, 19, 628–646.
B Struct. Crystallogr. Cryst. Chem. 1982, 38, 1584–1586. [110] J. T. Groves, Nat. Chem. 2014, 6, 89–91.
[74] T. T. Tsou, J. C. Huffman, J. K. Kochi, Inorg. Chem. 1979, 18, 2311– [111] K. Muto, J. Yamaguchi, K. Itami, J. Am. Chem. Soc. 2012, 134, 169–172.
2317. [112] Z. Li, S.-L. Zhang, Y. Fu, Q.-X. Guo, L. Liu, J. Am. Chem. Soc. 2009,
[75] R. Countryman, B. R. Penfold, J. Chem. Soc. D Chem. Commun. 1971, 131, 8815–8823.
1598. [113] I. Bach, R. Goddard, C. Kopiske, K. Seevogel, K.-R. Pörschke,
[76] M. Bühl, M. Håkansson, A. H. Mahmoudkhani, L. Öhrström, Organometallics 1999, 18, 10–20.
Organometallics 2000, 19, 5589–5596. [114] W. He, B. O. Patrick, P. Kennepohl, Nat. Commun. 2018, 9, 3866.
[77] P. N. Plessow, L. Weigel, R. Lindner, A. Schäfer, F. Rominger, M. [115] I. Bach, K.-R. Pörschke, R. Goddard, C. Kopiske, C. Krüger, A. Rufińska,
Limbach, P. Hofmann, Organometallics 2013, 32, 3327–3338. K. Seevogel, Organometallics 1996, 15, 4959–4966.
[78] J. J. Garcia, W. D. Jones, Organometallics 2000, 19, 5544–5545. [116] S. D. Ittel, H. Berke, H. Dietrich, J. Lambrecht, P. Härter, J. Opitz, W.
[79] J. J. Garcia, N. M. Brunkan, W. D. Jones, J. Am. Chem. Soc. 2002, 124, Springer, in Inorg. Synth. Vol 17 (Ed.: A.G. Macdiarmid), Wiley, 2007, pp.
9547–9555. 117–124.
[80] T. A. Ateşin, T. Li, S. Lachaize, W. W. Brennessel, J. J. García, W. D. [117] S. M. Webb, Phys. Scr. 2005, T115, 1011–1014.
Jones, J. Am. Chem. Soc. 2007, 129, 7562–7569. [118] B. Ravel, M. Newville, J. Synchrotron Radiat. 2005, 12, 537–541.
[81] D. J. Mindiola, R. Waterman, D. M. Jenkins, G. L. Hillhouse, Inorg. Chim. [119] M. U. Delgado-Jaime, P. Kennepohl, J. Synchrotron Radiat. 2010, 17,
Acta 2003, 345, 299–308. 119–28.
[82] R. Waterman, G. L. Hillhouse, J. Am. Chem. Soc. 2008, 130, 12628– [120] M. U. Delgado-Jaime, C. P. Mewis, P. Kennepohl, J. Synchrotron Radiat.
12629. 2010, 17, 132–7.
[83] J. J. Curley, K. D. Kitiachvili, R. Waterman, G. L. Hillhouse,
Organometallics 2009, 28, 2568–2571.
[84] D. J. Mindiola, R. Waterman, V. M. Iluc, T. R. Cundari, G. L. Hillhouse,
Inorg. Chem. 2014, 53, 13227–13238.
[85] R. Doi, K. Kikushima, M. Ohashi, S. Ogoshi, J. Am. Chem. Soc. 2015,
137, 3276–3282.
[86] J. E. Penner-Hahn, in Compr. Coord. Chem. II From Biol. to
Nanotechnol. (Eds.: A.B.P. Lever, J.A. McCleverty, T.J. Meyer), Elsevier
Pergamon, 2005, pp. 159–186.
[87] N. C. Tomson, K. D. Williams, X. Dai, S. Sproules, S. DeBeer, T. H.
Warren, K. Wieghardt, Chem. Sci. 2015, 6, 2474–2487.
[88] T. E. Westre, P. Kennepohl, J. G. DeWitt, B. Hedman, K. O. Hodgson,
E. I. Solomon, J. Am. Chem. Soc. 1997, 119, 6297–6314.