of 44
Electronic structures of nickel(II)-bis(indanyloxazoline)-dihalide
catalysts: Understanding ligand field contributions that promote
C(sp
2
)–C(sp
3
) cross-coupling
Brendon J. McNicholas
†,1
,
Z. Jaron Tong
†,2
,
Daniel Bím
1
,
Raymond F. Turro
2
,
Nathanael P.
Kazmierczak
1
,
Jakub Chalupský
3,4
,
Sarah E. Reisman
2
,
Ryan G. Hadt
1,*
1
Division of Chemistry and Chemical Engineering, Arthur Amos Noyes Laboratory of Chemical
Physics, California Institute of Technology, Pasadena, California 91125, United States
2
Division of Chemistry and Chemical Engineering, The Warren and Katherine Schlinger
Laboratory for Chemistry and Chemical Engineering, California Institute of Technology,
Pasadena, California 91125, United States
3
J. Heyrovský Institute of Physical Chemistry, The Czech Academy of Sciences, Dolejškova 3,
Prague 8, Czech Republic
4
Institute of Organic Chemistry and Biochemistry, Academy of Sciences of the Czech Republic,
Flemingovo náměstí 2, 166 10 Prague 6, Czech Republic
Abstract
Ni
II
(IB)
dihalide [
IB
= (3a
R
,3a’
R
,8a
S
,8a’
S
)-2,2’-(cyclopropane-1,1-diyl)bis(3a,8a-dihydro-8
H
-
indeno[1,2-
d
]-oxazole)] complexes are representative of a growing class of first-row transition
metal catalysts for the enantioselective reductive cross-coupling of C(sp
2
) and C(sp
3
)
electrophiles. Recent mechanistic studies highlight the complexity of these ground state cross-
couplings, but also illuminate new reactivity pathways stemming from one-electron redox
and their significant sensitivities to reaction conditions. For the first time, a diverse array of
spectroscopic methods coupled to electrochemistry has been applied to Ni
II
-based pre-catalysts
to evaluate specific ligand field effects governing key Ni-based redox potentials. We also
experimentally demonstrate DMA solvent coordination to catalytically-relevant Ni complexes.
Coordination is shown to favorably influence key redox-based reaction steps and prevent other
deleterious Ni-based equilibria. Combined with electronic structure calculations, we further
provide a direct correlation between reaction intermediate frontier molecular orbital energies
and cross-coupling yields. Considerations developed herein demonstrate the use of synergic
spectroscopic and electrochemical methods to provide concepts for catalyst ligand design and
rationalization of reaction condition optimization.
*
Corresponding Author: rghadt@caltech.edu.
Co-first author
Supporting Information
The Supporting Information is available free of charge on the ACS Publications website at DOI: Materials and methods, NMR spectra,
additional UV-vis-NIR, CD, and MCD spectra, additional voltammetry and spectroelectrochemistry data, X-ray crystallographic
parameters, and DFT and CASSCF/CASPT2 input parameters and results.
HHS Public Access
Author manuscript
Inorg Chem
. Author manuscript; available in PMC 2023 September 27.
Published in final edited form as:
Inorg Chem
. 2023 August 28; 62(34): 14010–14027. doi:10.1021/acs.inorgchem.3c02048.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Graphical Abstract
1. Introduction
The number of accessible oxidation states and the possibility of both one- and two-electron
redox reactions make nickel a well-suited alternative to precious metals such as iridium and
platinum for cross-coupling catalysis. Since the first report of reductive coupling of aryl
halides to biaryl products by bis(1,5-cyclooctadiene)nickel
0
in 1971,
1
numerous reports of
coupling reactions, including enantioselective cross-coupling, have appeared.
2
Early studies
used electrochemistry to render these reactions catalytic in nickel,
3
and in 2007 Durandetti
et al.
demonstrated that elemental manganese could be used as the terminal reductant for the
Ni(bipyridine)Br
2
-catalyzed reductive cross-coupling of iodobenzene with
α
-chloroesters in
up to 87% yield.
4
The first highly selective enantioconvergent reductive cross-coupling was
reported by one of our groups in 2013.
5
Using bis(4-phenyloxazoline) (PhBOX), elemental
manganese, and a 30% DMA/THF solvent mixture, high yields and enantioselectivity were
obtained for cross-coupling of acyl chlorides and benzyl chlorides. Following this, a variety
of chiral ligand frameworks were shown to be effective for the reductive cross-coupling
of different electrophile pairs.
2
For example, bis(indanyloxazoline) (Scheme 1, bottom,
IB
)
Ni complexes catalyze a variety of reductive alkenylation reactions, including formation
of enantioenriched allylic silanes and alkenes with aryl-substituted tertiary stereogenic
centers (Scheme 1, bottom).
6
,
7
More recently, one of our groups has also demonstrated
electrocatalytic competency of these Ni
II
complexes for cross-coupling reactions.
8
A fundamental and detailed description of both the ground- and excited-state electronic
properties of transition metal complexes can help infer catalytic operativity and competency.
For example, one of our groups has recently provided detailed electronic structural
and mechanistic studies of the ground- and excited-state properties of low-spin Ni
II
bipyridine aryl halide complexes and their Ni
I
photogenerated intermediates. These studies
identified key structure-function relationships relevant for excited state bond homolysis
and oxidative addition reactivity, which allowed for the mechanistic analysis of Ni
I
-
mediated activation of strong C(sp
2
)–Cl bonds.
9
,
10
Our groups also recently participated
in collaborative work that utilized cyclic voltammetry under catalytic conditions and UV-vis-
McNicholas et al.
Page 2
Inorg Chem
. Author manuscript; available in PMC 2023 September 27.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
NIR spectroelectrochemistry to better understand the interplay of Ni and Cr oxidation states
for Nozaki-Hiyama-Kishi coupling.
11
Another study by Neidig
et al.
utilized a combination
of Mössbauer spectroscopy, magnetic circular dichroism, and computations to demonstrate
that the kinetic competency of four-coordinate organoiron toward halogen abstraction is
dependent on the accessibility and relative energy of iron-based orbitals.
12
Despite the rapidly growing number of Ni-catalyzed reductive cross-coupling reactions
being developed, there have been relatively few detailed mechanistic studies, particularly
of bis(oxazoline) Ni complexes. Additionally, although solvent typically has a profound
influence on the yield and selectivity of these reactions,
13
18
the influence of solvent
coordination to the active catalyst has not been thoroughly studied. A previous study
has crystallographically characterized a five-coordinate, DMSO-bound Ni
II
phenanthroline
complex, also suggesting the potential influence of solvent in the catalytic cycle.
19
A recent
study by Diao
et al.
has suggested a Ni
I/III
redox cycle and radical formation and capture
are pertinent for catalysis by (biOx)Ni
II
ArX catalysts, which revised previous suggestions of
involvement of Ni
0/II
oxidation states.
20
In general, uncertainty still exists regarding how the
specific ligand, as well as the potential role of solvent coordination, influence both low-spin
and high-spin Ni
II
cross-coupling catalysis.
To this end, we provide the first comprehensive spectroscopic and electrochemical
investigation of two reductive cross-coupling catalysts,
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
[
IB
= (3a
R
,3a’
R
,8a
S
,8a’
S
)-2,2’-(cyclopropane-1,1-diyl)bis(3a,8a-dihydro-8
H
-indeno[1,2-
d
]-oxazole)], to provide broad insight into the complex conditions and mechanism (Scheme
1). Although studies related to these complexes by one of our groups were primarily
in the context of developing reductive alkenylation, we note that these and similar
complexes have also been reported for a number of other reactions, including reductive
arylation, 1,2-alkynylboration, and 1,2-vinylboration, further motivating their detailed
experimental and computational investigation.
21
31
Here we use variable-temperature (VT)
UV-vis-NIR absorption, circular dichroism (CD), vibrational CD (VCD), and magnetic
CD (MCD) spectroscopies, coupled with cyclic voltammetry, spectroelectrochemistry, and
DFT/TDDFT and multi-reference (CASSCF/CASPT2) calculations to elucidate specific
electronic structure contributions to reactivity, as well as the influence of solvent on the
efficacy of catalytic transformations (Scheme 1). Quantifying halide-dependent spectral
features directly connects ligand field strength and redox potentials. Similarly, studying
three solvents with different donor numbers (DNs) provides a route toward understanding
previously reported empirical solvent optimization studies. We demonstrate that DMA
solvent can coordinate to catalytically-relevant Ni species and evaluate the potential role
of solvent coordination in their reactivity. Based on previously reported yields and findings
reported here, we suggest solvent coordination and low temperature can favorably influence
driving forces and kinetic barriers of key reaction steps leading to cross-coupled product.
These studies provide new insights relevant to the catalytic reactivity of chiral bis(oxazoline)
Ni catalysts that have been recently popularized for a variety of asymmetric Ni-catalyzed
cross-coupling reactions (Scheme 1).
McNicholas et al.
Page 3
Inorg Chem
. Author manuscript; available in PMC 2023 September 27.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
2. Results and Analysis
2.1 Room-Temperature UV-Vis-NIR, CD, and MCD Characterization of Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
in a Non-Coordinating Solvent
Due to the absence of low-energy charge transfer bands in the UV-vis-NIR, electronic
absorption spectra of
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
in DCM (donor number, DN = 2.4)
32
,
33
(Figures 1A–C and 1D–F, respectively) provide rich insights into the Ni
II
ligand field
through observation of numerous spin-allowed and spin-forbidden ligand field transitions
(Table 1). These transitions arise from the orbital triplet ground state (i.e.,
3
T
1
(F) in an
idealized
T
d
geometry) of the S = 1
d
8
Ni
II
complexes and exhibit correspondingly weak
molar absorptivities at 294 K. CD and MCD spectroscopies provide complementary, signed,
spectroscopic methods to resolve overlapping ligand field transitions. Due to the presence of
the chiral
IB
ligand, the Ni
II
-based ligand field transitions exhibit CD intensity (Figures 1B
and 1E), and both complexes exhibit room temperature MCD signals at 1.4 T (Figures 1C
and 1F, respectively). A detailed discussion of band assignments can be found in Supporting
Information Section S3.
The ligand field transitions of
Ni
II
(IB)Cl
2
relative to
Ni
II
(IB)Br
2
are blueshifted by ~120 –
710 cm
−1
in DCM (overlaps in Figure S17), consistent with stronger donation from chloride
relative to bromide. To estimate the relative ligand field strengths, we average the assigned
spin-allowed transitions (bands 1 – 7). Doing so provides relative ligand field strengths
of ~12 365 cm
−1
and ~11 975 cm
−1
(Δ = ~390 cm
−1
) for
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
,
respectively, in accord with the greater ligand field strength in the chloride complex relative
to bromide. As will be shown below, the ligand field bands and ligand field strengths can be
directly related to the energies of the Ni
II
-based redox active molecular orbitals (RAMOs)
and, thus, complex redox potentials (
vide infra
, Section 3.3). To the best of our knowledge,
this provides the first experimental connection between ligand field spectroscopy and the
electrochemical potentials of Ni
II
-based enantioselective cross-coupling catalysts.
2.2 Vibrational CD Spectroscopy of Ni
II
(IB)Cl
2
and Ni
II
(IB)Br
2
Pseudo-
T
d
Ni(II)/Co(II) and pseudo-
O
h
V(III) complexes can exhibit large splittings of
their orbital triplet ground states. These splittings can be observed using techniques such as
electronic Raman or VCD. The latter is the infrared analogue of electronic CD spectroscopy
and provides a means to determine absolute stereochemical configurations or probe low-
energy electronic transitions in chiral transition metal complexes.
34
,
35
VCD signals are
observed for both
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
complexes dissolved in
d
2
-DCM (Figure 2).
Note high sample concentrations are necessary for VCD measurements. Despite substantial
trimerization of
Ni
II
(IB)Cl
2
at these concentrations (
vide infra
, Section 2.4), the observed
VCD transition is still assigned to the four-coordinate species, as evidenced by calculations
(Figure S121) and spectral consistency with the
Ni
II
(IB)Br
2
analogue, which does not
undergo trimerization at high concentrations in DCM. Furthermore, to confirm the observed
VCD spectral intensity for
Ni
II
(IB)X
2
complexes corresponds to a ligand field transition, the
d
10
complex,
Zn
II
(IB)Cl
2
, was synthesized in a manner analogous to the Ni
II
complexes.
Note that an analogous complex, Cu
II
(IB)Cl
2
, has been previously reported.
36
The VCD
spectrum of
Zn
II
(IB)Cl
2
in
d
2
-DCM does not exhibit electronic absorption in the region
McNicholas et al.
Page 4
Inorg Chem
. Author manuscript; available in PMC 2023 September 27.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
of 1800 – 3200 cm
−1
(Figure 2).
Ni
II
(IB)Cl
2
exhibits a moderately sharp (FWHM =
~730 cm
−1
) ligand field transition at ~2170 cm
−1
, while
Ni
II
(IB)Br
2
exhibits a transition
(FWHM = ~970 cm
−1
) at ~2210 cm
−1
. For both complexes, this band (band 1,
3
B
1
(F)
3
A
2
transition in
C
2v
) arises from a transition within the orbital triplet ground state
(
3
T
1
(F) in
T
d
), which is split by low symmetry distortions and spin-orbit coupling. Note the
positive sign of the VCD band is consistent with the other
3
A
2
states observed with CD
(Figure 1B and 1E). While the relative ground state splittings are quite similar between the
two complexes, the small increase in splitting for
Ni
II
(IB)Br
2
is likely due to the greater
spin-orbit coupling constant for Br relative to Cl.
This application of VCD determines electronic excited state energies of the low-symmetry
split orbital triplet ground state of transition metal-based enantioselective cross-coupling
catalysts for the first time. These data, combined with the UV-vis-NIR, electronic CD, and
MCD data, have allowed for the experimental determination of a complete ligand field
energy level diagram for the
Ni
II
(IB)X
2
complexes. All transitions to individual excited
states in
T
d
and
C
2v
symmetry are assigned in the correlation diagram in Figure 3.
2.3 Room Temperature UV-Vis-NIR, CD, and MCD Characterization of Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
in Coordinating Solvents
To complement data obtained in DCM (poor
σ
and
π
donor) and to assess the solvent-
dependent behavior of
Ni
II
(IB)X
2
complexes, room temperature UV-vis-NIR absorption,
CD, and 1.4 T MCD spectra were also acquired in MeCN (moderate
σ
donor and moderate
π
acceptor) and DMA (very strong
σ
donor and moderate
π
acceptor). These solvents were
selected for their relative differences in dielectric constant and Lewis basicity/acidity to
assess the effects of solvent environment on geometric and electronic structure, which can
subsequently be correlated with catalytic activity and selectivity.
UV-vis-NIR spectra of
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
exhibit significant solvent dependence
(Figures S9 and S10). In addition to decreased overall intensities compared to transitions
observed in DCM, new spectral intensity grows in at ~23 000 cm
−1
, with weak intensity
in MeCN and greatest intensity in DMA. These spectral changes are also manifested in
the solvent-dependent CD and MCD spectra (Figures S44–S47), with the new spectral
intensity at ~23 000 cm
−1
corresponding to a new negative band in CD and MCD. As
demonstrated in Section 2.4 below, this new band reflects an equilibrium between the four-
and five-coordinate, solvent coordinated species. Thus, electronic spectroscopies provide a
direct handle on DMA coordination to Ni
II
dihalide complexes relevant to catalysis.
2.4 Variable-Concentration and Variable-Temperature UV-Vis-NIR Spectroscopy of
Ni
II
(IB)Cl
2
and Ni
II
(IB)Br
2
Solvent and additive evaluation are necessary steps in the optimization of transition metal-
catalyzed organic reactions. However, these steps can be somewhat arbitrary and rely on
a large empirical screening matrix. Previous studies of
Ni
II
(IB)X
2
demonstrated catalytic
yields are maximized in DMA and when the reaction is cooled to 0–5 °C.
6
8
By obtaining
VT UV-vis-NIR spectra for both
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
in DCM and DMA, we
McNicholas et al.
Page 5
Inorg Chem
. Author manuscript; available in PMC 2023 September 27.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
aimed to provide insight into precatalyst speciation and, in turn, to provide experimental
thermodynamic data for rational catalyst and condition design.
As the concentration of
Ni
II
(IB)Cl
2
is increased in DCM, a noticeable color change from
pink to orange is observed. Correspondingly, new electronic absorption bands are observed
at 13 100 cm
−1
and 22 840 cm
−1
with increasing concentration (Figure S16). These
additional absorption bands are ascribed to the formation of a [
Ni
II
(IB)Cl
2
]
3
μ-Cl
trimer.
A previously obtained crystal structure of this species shows it possesses both five- and
six-coordinate formal Ni
II
centers.
37
The UV-vis-NIR spectra of
Ni
II
(IB)Cl
2
in DCM (4.3 mM) and DMA (3.6 mM) also
depend on temperature (Figure 4A and 4B). Because of the number of overlapping
transitions present in each spectrum, the VT spectra were resolved using global modeling
of the temperature dependence through nonlinear regression at multiple wavelengths
and bootstrapping (Figures S13–S15).
38
,
39
In DMA, the VT UV-vis-NIR spectra reflect
an equilibrium between the four- and five-coordinate, DMA coordinated species for
Ni
II
(IB)Cl
2
(five-coordinate absorption maximum at 22 840 cm
−1
) and
Ni
II
(IB)Br
2
(five-
coordinate absorption maximum at 23 230 cm
−1
). In DCM, the VT UV-vis-NIR spectra
reflect an equilibrium between a monomeric and trimeric form (trimer absorption maxima
at 22 840 cm
−1
and 20 160 cm
−1
). We note the excellent agreement between the
resolved spectra of the four-coordinate species and the spectra of isolated
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
obtained in DCM (Figure 4 and S12). In contrast to
Ni
II
(IB)Cl
2
,
Ni
II
(IB)Br
2
exhibited no trimerization up to 152.1 mM or in VT studies down to –85 °C (Figures S18–
S19).
In addition to resolving the spectra for four-coordinate, five-coordinate, and trimeric species,
these fits provide thermodynamic parameters based on the two equilibrium expressions,
K
eq
=
Ni
IB
X
2
3
Ni
IB
X
2
3
,
K
eq
=
Ni
Solv
.
IB
X
2
Ni
IB
X
2
(1)
K
eq
values for DMA coordination to
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
are similar (0.58 and 0.97,
respectively, Table 2). Furthermore,
K
eq
= 16.7
is calculated at 294 K for formation of the
[
Ni
II
(IB)Cl
2
]
3
μ-Cl
trimer in DCM. Despite the exergonicity of this process (Table 2), the
extent of reaction is only 0.09% at 4.3 mM, which is why essentially no trimer is observed
at room temperature at low Ni
II
concentrations. As expected for associative reactions, all
entropy values for solvent coordination and trimerization are negative. Note that the extent
of temperature-dependent reaction is primarily sensitive to the ratio of the enthalpy to
entropy, and the global fitting can extract this value with <1% uncertainty. However, global
fitting can additionally extract the absolute standard enthalpies and entropies, albeit with
higher uncertainty. From
K
eq
values, we estimate concentrations of ~37% and ~49% for
DMA-coordinated
Ni
II
(IB)Cl
2
(3.6 mM) and
Ni
II
(IB)Br
2
(3.5 mM), respectively, at 294 K.
Upon cooling to 273 K, these values increase significantly to ~53% and ~69%, respectively.
McNicholas et al.
Page 6
Inorg Chem
. Author manuscript; available in PMC 2023 September 27.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
As depicted in Figure 5, the preceding analysis provides the first detailed view of the effects
of dielectric constant, concentration, and solvent donicity on the speciation equilibria of
Ni
II
precatalyst solutions. With high solvent donicity, the equilibrium shifts toward a five-
coordinate, solvent coordinated species for both
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
. Increasing
concentration and low dielectric constant shifts the equilibrium toward a trimeric species
for
Ni
II
(IB)Cl
2
. Thus, solvent coordination, temperature, and catalyst concentration are
important effects that contribute to catalyst speciation and activity under reaction conditions
(
vide infra
,
Discussion
). VT spectroscopies provide a direct handle on the speciation and the
associated thermodynamics.
2.5 Cyclic Voltammetry of Ni
II
(IB)Cl
2
and Ni
II
(IB)Br
2
in DCM, MeCN, and DMA
To assess the effects of solvent donicity and dielectric constant on the electrochemical
properties of
Ni
II
(IB)X
2
complexes, scan rate-dependent cyclic voltammetry data were
acquired in DCM, MeCN, and DMA (Figure 6 and Figures S48–S59). Table 3 provides
peak and formal potentials for initial redox events, while Table S1 provides peak and
formal potentials for unique re-oxidation and re-reduction events that result from chemical
reactions following initial electron transfers. Additional details and discussions are provided
in Supporting Information Section S5.
Previous studies have provided formal potentials for both aromatic and non-aromatic Ni
II
diimine systems, with many reports providing kinetic analyses with substrate present.
41
,
20
However, to our knowledge, this is the first example of detailed solvent-dependent
electroanalytical chemistry for non-aromatic Ni
II
cross-coupling catalysts. In general,
precatalyst electrochemical responses are remarkably solvent dependent. In all three
solvents,
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
both exhibit single, electrochemically irreversible
reduction events with significantly shifted oxidative waves (Figures 6, and S50, S54, S58).
The general irreversibility required use of peak potentials (
E
p,a
or
E
p,c
), potentials at half of
the peak current value (
E
p/2
), and inflection potentials (accurate estimate of formal potential,
E
0’
) for analysis.
42
Our measured Fc formal potentials in DMA and MeCN are 85 and 91
mV vs 0.01 M Ag
+/0
, respectively, indicating accurate conclusions can be drawn regarding
solvent effects on measured formal potentials of the Ni complexes in these solvents. In
DCM, the measured Fc formal potential is 215 mV. Therefore, measured formal potentials in
DCM will appear negatively shifted relative to values in DMA and MeCN.
Based on shifts in peak potential as a function of scan rate and scan rate normalized
voltammetry (current function) in all three solvents (Figures S48–S58), as well as
differential pulse voltammetry and variable temperature voltammetry in DMA for
Ni
II
(IB)Cl
2
(Figure S58), we can draw some insightful conclusions regarding the reduction
mechanism, as the current function and shift in peak potential are dictated by the particular
chemical and electrochemical mechanism. These conclusions also apply to
Ni
II
(IB)Br
2
. We
ascribe the reduction of both complexes to a concerted E
q
C
i
(in DMA and MeCN, solvent
coordination and/or halide loss occur in concert) or step-wise E
i
C
i
mechanism (DCM),
where slow electron transfer is followed by rapid halide loss. In DCM, three-coordinate
Ni
I
(IB)X
will be generated upon reduction, with no subsequent solvent coordination. The
lack of return current, shift in peak potential as a function of the logarithm of the scan
McNicholas et al.
Page 7
Inorg Chem
. Author manuscript; available in PMC 2023 September 27.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
rate near 29.6 mV, with ~33 mV observed here, and decrease in the current function
toward a limiting value as the scan rate is increased supports a kinetically-controlled,
stepwise reduction followed by rapid halide loss (Figures S48 and S50). Activation of
DCM by other nickel complexes supported by naphthyridine-diimine ligands has been
observed previously.
43
However, spectroelectrochemical data obtained in DCM do not
support reactivity of the
Ni
I
(IB)X
with solvent (Figures S73–S75). Overall, this analysis
featuring electron transfer coupled to rapid halide loss is consistent with halide dissociation
observed previously using extended X-ray absorption fine structure (EXAFS) for a low-spin
Ni
II
biOx aryl halide complex upon reduction with potassium graphite.
20
Experimental formal potentials for chemically-coupled reduction of
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
to
Ni
I
(IB)Cl
and
Ni
I
(IB)Br
in DCM are −1.47 V and −1.26 V vs. Fc
+/0
,
respectively (Table 3). It is therefore ~0.21 V (~1695 cm
−1
) harder to reduce
Ni
II
(IB)Cl
2
relative to
Ni
II
(IB)Br
2
. This observation is consistent with the energetic shifts in the spin-
allowed ligand field bands in DCM in experiment (
vide supra
, Section 2.1) and calculations
(
vide infra
, Section 3.2 and 3.3). In MeCN and DMA,
Ni
II
(IB)Cl
2
(Figures S54 and S58)
and
Ni
II
(IB)Br
2
(Figure 6, left) both exhibit superficially quasi-reversible voltammetry
for the reduction. Experimental formal potentials for chemically-coupled reduction in
MeCN/DMA of
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
are −1.32/−1.47 V and −1.05/−1.23 V vs.
Fc
+/0
, respectively (Δ = ~0.27/~0.24 V (~2180/1935 cm
−1
)). Thus, for all solvents used here,
it is harder to reduce
Ni
II
(IB)Cl
2
relative to
Ni
II
(IB)Br
2
.
Based on the VT UV-vis-NIR data in DMA (
vide supra
, Section 2.4), both the Ni
II
four-coordinate and five-coordinate solvent adducts exist in equilibrium, and this can
potentially influence the electrochemistry measured in this solvent. One possibility for the
reduction mechanism for these species is reduction followed by halide loss and, for the
four-coordinate portion of the complex, coordination of DMA to the Ni
I
center, which
could occur in a concerted or stepwise fashion. For a concerted mechanism, the anticipated
shift in peak potential as a function of log(
v
) is 29.6/
α
mV, where
α
is the transfer
coefficient for electron transfer.
42
Based on the observation of only one differential pulsed
voltammetry wave on the forward scan and the shift in peak potential with log(
v
) (~77–
104 mV), we propose that the reduction and chemical follow up reaction in both MeCN
and DMA (i.e., solvent coordination at Ni
I
) is a concerted process. The two return waves
observed scanning oxidatively suggest generation of a halide-dissociated species that is
re-oxidized at more positive potentials. This conclusion is supported by VT differential
pulse voltammetry (Figure S58), where the differential current at the more positive wave
decreases as temperature is decreased, while the differential current at the wave ascribed
to re-oxidation of five-coordinate Ni
I
increases. Based on behavior previously observed
for these systems and our computed formal potentials,
20
the more positive re-oxidation
could arise from re-oxidation of a Ni
I
/Ni
I
dimer that forms after the initial reduction.
However, we favor the interpretation featuring re-oxidation of the halide-dissociated species
based on computed formal potentials (
vide infra
, Section 3.3) and lack of return oxidation
near the reduction event in DCM, where Ni
I
is anticipated to dimerize rapidly. Further
supporting our hypothesis, an additional wave near where three-coordinate Ni
I
is predicted
to oxidize is present in MeCN, but not in DMA (Figures S52 and S54), supporting the
McNicholas et al.
Page 8
Inorg Chem
. Author manuscript; available in PMC 2023 September 27.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
weaker coordination affinity of MeCN and our assignment of the species generated upon
reduction.
Potentials for chemically-coupled reductions in DMA are more negative relative to MeCN
by ~200 mV for both complexes. As discussed further in Section 3.3, this difference is
ascribed to DMA being a higher donicity solvent and coordinating to the Ni
II
center. Note
that solvent coordination is not observed in DCM and only weakly so in MeCN. It is further
interesting to note that the reduction potential for
Ni
II
(IB)Cl
2
in both DCM and DMA is
−1.47 V vs. Fc
+/0
, respectively; for
Ni
II
(IB)Br
2
, these are −1.26 V and −1.23 V vs. Fc
+/0
,
respectively. The similarity in reduction potentials in DCM and DMA is ascribed to the
relative Fc formal potentials in DCM vs. DMA and the role of solvent in facilitating the
Ni–X bond rupture upon one-electron reduction, with the anionic halide loss more facile
in DMA relative to DCM. Because of these considerations and the electronic structure
calculations presented in Section 3.3, the more quantitative comparison of potentials for
the reduction with and without coordinated solvent is that between MeCN and DMA.
Furthermore, the temperature-dependent cyclic voltammetry demonstrates a negatively
shifted reduction potential as the temperature is lowered, which may be due to increasing
the relative amount of five-coordinate species. Thus, overall, solvent coordination results in
a harder to reduce Ni
II
center. By extension, this can be further translated to a more reducing
Ni
I
species, which, under catalytic conditions, can facilitate oxidative addition (
vide infra
,
Discussion
).
44
Ligand field and bonding effects on Ni
II
-based redox potentials can be further elucidated
using electronic structure calculations (
vide infra
, Section 3.3) and through correlations
to electronic spectroscopy, as transitions to the RAMO are also observed experimentally.
Differences in measured redox potentials correlate directly with specific structural
influences on the energy of the RAMO.
Finally, based on the measured formal potentials, proposed electrochemical mechanisms,
and additional electronic structure calculations of redox potentials (
vide infra
, Section
3.3), we do not believe
Ni
0
(IB)X
2
(or Ni
0
in any form) is thermodynamically accessible
in the electrochemical window of common electrochemistry solvents, which supports a
Ni
I/III
catalytic cycle for reductive alkenylation and potentially related reactions involving
bis(oxazoline)–Ni complexes.
20
No additional reduction beyond Ni
I
is required for oxidative
addition of substrates for which this catalyst has been previously demonstrated to be
competent, and this finding has important mechanistic implications for bis(oxazoline)–Ni-
catalyzed reactions more generally.
2.6 Spectroelectrochemistry of Ni
II
(IB)Cl
2
and Ni
II
(IB)Br
2
in DCM and DMA
To rationalize the noticeable difference in electrochemical response of
Ni
II
(IB)X
2
in
DCM vs. DMA and to further understand the solvent-dependent catalytic activity,
time-based spectroelectrochemical measurements were performed for both negative and
positive polarizations. These measurements are the first comprehensive solvent-dependent
spectroelectrochemistry for Ni
II
cross-coupling catalysts. All experimental plots are shown
in Figures S61–S87. In conjunction with calculations (
vide infra
, Figures S133–S140), we
can assign transient spectra to possible species generated under polarized conditions. In
McNicholas et al.
Page 9
Inorg Chem
. Author manuscript; available in PMC 2023 September 27.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
DCM, polarization negative of the first reduction generates a spectrum with a slightly blue-
shifted, higher-energy ligand field transition, consistent with calculated spectra for
Ni
I
(IB)X
(Figures S133 and S135).
Reduction of both
Ni
II
(IB)Br
2
and
Ni
II
(IB)Cl
2
in DMA generates spectra consistent with
calculated spectra corresponding to a solvent-coordinated Ni
I
species,
Ni
I
(IB)(DMA)X
(Figures S134 and S136). All ligand field bands decay in intensity upon oxidation, which
is ascribed to oxidative degradation or oligomerization. Spectra obtained after positive
polarization of
Ni
II
(IB)Br
2
are consistent with bromide speciation.
45
3. Computational Results
In this section, we sought to gain further insights into the electronic structures of the
precatalysts by comparing experimental spectra and redox potentials with computed ground-
and excited-state properties obtained from a combination of DFT, time-dependent DFT
(TDDFT), and multireference CASSCF/CASPT2 calculations. TDDFT and multireference
methods predict electronic transitions below ~25 000 cm
−1
are ligand field excitations,
and differences between computed electronic spectra of
Ni
II
(IB)X
2
complexes arise from
differences in ligand field strength. Generally, increased ligand field strength destabilizes the
β
lowest unoccupied molecular orbitals (
β
-LUMOs), which negatively shifts the reduction
potential of
Ni
II
(IB)Cl
2
relative to
Ni
II
(IB)Br
2
. DFT calculations also corroborate that,
depending on reaction conditions (such as the choice of coordinating/non-coordinating
solvent or Ni
II
concentration), precatalysts can exist in complex equilibria featuring four-
coordinate
Ni
II
(IB)X
2
, five-coordinate
Ni
II
(IB)(solv.)X
2
, or trimeric [
Ni
II
(IB)X
2
]
3
μ-X
species. Finally, we demonstrate the role of ligand–metal covalency in tuning the relative
reactivity of the catalyst resting state; the bonding of this species further manifests in a
correlation between reaction yield and its oxidation potential relative to those of C(sp
3
)
radicals.
3.1 DFT and TDDFT Calculated Thermodynamics and Spectra of Ni
II
(IB)Cl
2
and Ni
II
(IB)Br
2
Precatalysts
Using a TPSSh functional with a conductor-like polarizable continuum model (CPCM)
(see Computational Details in Section S8 of the Supporting Information), the ground state
wave functions of
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
are high-spin (
S
= 1) Ni
II
. TDDFT was
used to calculate the electronic transition energies (Table 4). The intensities of the lower-
energy ligand field bands are underestimated relative to experiment and do not contribute
significantly to the overall predicted spectra (Figures S92–S93). Additionally, two-electron
excitations and spin-flip transitions are inaccessible through conventional TDDFT and,
thus, are not observed in the spectral predictions (e.g.,
3
A
2
(F), band 4 and
1
T
2
/
1
E, band
i excited states in
T
d
). For
d
8
Ni
II
, there are only six spin-allowed ligand field transitions
accessible using TDDFT (Table 4). Therefore, full assignment of all absorption bands
from Section 2.1 cannot be achieved using this approach. A more detailed analysis must
be obtained from the multi-reference CASSCF/CASPT2 calculations (
vide infra
, Section
3.2). Nonetheless, qualitative correlations can be made by comparing to assignments in
the parent
T
d
point group (Table 4). While the absolute calculated energies of the ligand
McNicholas et al.
Page 10
Inorg Chem
. Author manuscript; available in PMC 2023 September 27.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
field transitions are not well-reproduced with TDDFT, the average energies agree well with
the experimentally determined relative ligand field strengths (
vide supra
, Section 2.1). For
example, experimental values of ~12 365 cm
−1
and ~11 975 cm
−1
(Δ = ~390 cm
−1
) were
determined for
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
, respectively, and are computationally estimated
to be ~13 780 cm
−1
and ~13 370 cm
−1
(Δ = ~410 cm
−1
).
In addition to correlating TDDFT calculations to experimental precatalyst spectra, they can
be further utilized to understand the equilibria discussed in Section 2.4 (see Supporting
Information Section S8.1). The calculations corroborate that precatalysts can exist in
complex equilibria featuring four-coordinate
Ni
II
(IB)X
2
, five-coordinate
Ni
II
(IB)(solv.)X
2
,
or trimeric
[Ni
II
(IB)X
2
]
3
μ-X
species. Consistent with experiment, the calculated spectra
for five-coordinate species and trimeric species exhibit a significant blue shift for the most
intense calculated ligand field band (band 7; Figures S130–S132).
3.2
Ab initio
Multireference Calculations of the Ni
II
(IB)Cl
2
and Ni
II
(IB)Br
2
Precatalysts
Due to the inherent complications with TDDFT described above, we have also used
ab
initio
multireference calculations to compute and assign the experimentally observed ligand
field excitations. First, we have systematically probed the effects of active space variation
in CASSCF calculations (see Computational Details in Section S8 of the Supporting
Information) on the qualitative agreement of
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
with experimental
spectra. Regardless of active space size (Tables S24–S39), the ground state is exclusively
high spin. Notably, the largest active space used in this work, 22e,12o (five Ni 3
d
orbitals,
six halide 2
p
/3
p
orbitals, and the Ni(IB)
σ
bonding orbital), results in a single-reference
ground-state solution, with the highest weight of a single configuration in the CI vector
of ~95% for both
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
(Tables S29 and S35). This configuration
corresponds to an
S
= 1 triplet ground state with unpaired electrons in the
d
(x
2
-y
2
) and
d
(xz) orbitals (i.e., the same configuration as obtained from DFT calculations;
cf.
Figure
7). We note that inclusion of the occupied halide 2
p
/3
p
orbitals and the Ni(IB)
σ
bonding
orbital in the active space was essential to reproduce the experimental spectra. With this
optimized active space, UV-vis-NIR absorption, CD, and MCD spectra of
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
were calculated (Figure 8 and S141–S142). These calculations generally
support assignments of experimental data given in Section 2.1 (Table 4). Individual states
can be assigned based on the configuration state function with the largest weight in the CI
vector, in conjunction with the location of the 3
d
holes (see Supporting Information Section
S8.2).
Calculated signs for CD and MCD transitions are also consistent with experiment. All
3
A
2
excited states exhibit experimental and calculated positive CD bands and negative MCD
bands. The
3
A
1
excited state (band 3 in experiment) exhibits negative bands for both CD
and MCD. As supported by theory, the
3
B
1
excited state exhibits a negative CD band
and a positive MCD band. In contrast, the predicted sign for CD does not match that
observed experimentally for the higher-energy
3
B
2
excited state, though this is likely due
to this transition being formally forbidden and gaining intensity through either spin-orbit or
vibronic coupling, which can give either positive or negative differential intensity in CD.
46
The overall success of the CASSCF/CASPT2 approach in calculating UV-vis-NIR, and in
McNicholas et al.
Page 11
Inorg Chem
. Author manuscript; available in PMC 2023 September 27.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
particular CD/MCD spectra, is especially encouraging for future analyses of experimental
spectra for other chiral first row transition metal cross-coupling catalysts.
3.3 Computed Electrochemical Properties
Since the CASSCF CI vector indicates that the ground-state solutions are single-referent, we
can use DFT to interpret the effects of different halide ligands on the Ni
III/II
, Ni
II/I
, and Ni
I/0
reduction potentials of the
Ni
II
(IB)X
2
complexes and connect them to thermodynamically
accessible redox pathways of Ni-based reductive cross-coupling catalysis. Table 5 provides
computed potentials for various electrochemical processes (
vide supra
, Section 2.5).
Experimentally, the reduction of
Ni
II
(IB)X
2
complexes is chemically irreversible due to
halide dissociation. DFT calculations predict slightly positive halide dissociation energies of
Δ
G
dissoc
(CPCM) of ~6 kcal mol
−1
and ~7 kcal mol
−1
for Cl and Br
Ni
I
(IB)X
2
complexes,
respectively. The positive Δ
G
dissoc
may be associated with inaccuracies in halide solvation
energy when using a simple CPCM. The high electrolyte concentration in electrochemical
experiments, which is not accounted for in the computations, may also further shift the
equilibrium toward dissociation. For comparison, the calculated Δ
G
dissoc
of halide loss from
Ni
II
(IB)X
2
is significantly higher in energy (Δ
G
dissoc
(CPCM) = ~30 kcal mol
−1
and ~29
kcal mol
−1
for Cl and Br, respectively), indicating much stronger ligand–metal bonds for the
Ni
II
species.
The computed one-electron reduction potentials coupled to halide loss are −1.49 V and
−1.39 V vs. Fc
+/0
for
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
, respectively, with no additional
coordinated solvent ligand (Table 5). The absolute calculated values compare well with
those measured experimentally (Table 3), as does the calculated potential difference between
the two complexes (Δ = ~0.10 V (~805 cm
−1
) (calculated) vs. Δ = ~0.24 V (~1960 cm
−1
)
(experiment – average for all three solvents)). Thus, DFT calculations support the idea
that stronger ligand fields generally lead to a more negative potential for Ni
I
formation. In
that regard, more facile reduction of
Ni
II
(IB)Br
2
is attributed to the lower energy of the
β
-LUMOs by ~0.2 eV (Figure 7).
The computed one-electron reduction potentials coupled to halide loss also shift negatively
upon solvent coordination. For the Cl complex, calculated values are −1.49 V, −1.65 V, and
−1.66 V vs. Fc
+/0
for no solvent, MeCN, and DMA coordination, respectively. For the Br
complex, analogous calculated values are −1.39 V, −1.48 V, and −1.44 V vs. Fc
+/0
. These
negative potential shifts due to solvent coordination are also ascribed to modifications of
the Ni
II
β
-LUMOs. For example, average energy destabilizations of 0.40 eV and 0.47 eV
are observed for
Ni
II
(IB)Cl
2
and
Ni
II
(IB)Br
2
complexes, respectively, upon MeCN or DMA
coordination. Together, these calculations and the ligand field spectroscopy both indicate for
the first time that the RAMO energy is an excellent descriptor of the redox properties of
these metal-based cross-coupling catalysts, and DFT calculations provide a useful approach
for analyzing ligand contributions to potentials.
Finally, further extending these calculations to additional species, the computed reduction
of
Ni
I
(IB)X
to
Ni
0
(IB)X
or
Ni
0
(IB)
is calculated to occur at exceedingly negative formal
potentials (~−3 V to −4 V). These calculated potentials for Ni
0
formation are significantly
more negative than the reduction potentials of typical reductants used in cross-coupling
McNicholas et al.
Page 12
Inorg Chem
. Author manuscript; available in PMC 2023 September 27.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript