Polyannulated Bis(N-heterocyclic carbene)palladium Pincer Complexes for Electrocatalytic CO2 Reduction

Article abstract of DOI:10.1021/acs.inorgchem.5b01698

Phenanthro- and pyreno-annulated N-heterocyclic carbenes (NHCs) have been incorporated into lutidine-linked bis-NHC Pd pincer complexes to investigate the effect of these polyannulated NHCs on the ability of the complexes to electrochemically reduce CO

Full text of DOI:10.1021/acs.inorgchem.5b01698

Article
pubs.acs.org/IC
Polyannulated Bis(N-heterocyclic carbene)palladium Pincer
Complexes for Electrocatalytic CO2 Reduction
Jeffrey A. Therrien, Michael O. Wolf,* and Brian O. Patrick
Department of Chemistry, University of British Columbia, Vancouver, British Columbia V6T 1Z1, Canada
S
* Supporting Information
ABSTRACT: Phenanthro- and pyreno-annulated N-heterocyclic
carbenes (NHCs) have been incorporated into lutidine-linked bis-
NHC Pd pincer complexes to investigate the effect of these
polyannulated NHCs on the ability of the complexes to electro-
chemically reduce CO2 to CO in the presence of 2,2,2-trifluoro-
acetic acid and 2,2,2-trifluoroethanol as proton sources. These
complexes are screened for their ability to reduce CO2 and modeled
using density functional theory calculations, where the annulated
phenanthrene and pyrene moieties are shown to be additional sites
for redox activity in the pincer ligand, enabling increased electron
donation. Electrochemical and computational studies are used to
gain an understanding of the chemical significance of redox events
for complexes of this type, highlighting the importance of anion binding and dissociation.
donor type, bite angle, and changes in the redox potential of the
complex. These factors can act to modify the basicity of a
catalytic complex and the energy and stability of hydride
formation, leading to subsequent hydrogen production.7,8 In
general, less negative reduction potentials have been found to
be beneficial for improved selectivity for CO2 reduction.
High selectivity for CO2 reduction can also be achieved at
more reducing potentials if the active electrocatalyst species can
activate CO2 to a sufficiently high degree, resulting in negative
charge formation on the oxygen atoms of the bound CO2 and
higher basicity than the metal center. Density functional theory
(DFT) analysis of the bonding energy and degree of CO2
activation for the electrocatalysts Re(bpy)(CO)3Cl and [Pd-
(triphosphine)(CH3CN)]2+ show energetically favorable bind-
ing and activation of CO2 with O−C−O bond angles of
approximately 123° and 130°, respectively, at their energy
minima, which allow the use of water and HBF4, respectively, as
weakly and strongly acidic proton sources.6,9−11 The greater
degree of CO2 activation achieved with Re(bpy)(CO)3Cl
allows the use of weak acid sources at highly reducing potentials
while avoiding H2 production because the half-cell potential for
H+ reduction becomes more negative with increasing pKa.12
Conversely, a lower degree of CO2 activation requires the use
of stronger, less discriminating Brønsted acids for protonation
of the bound CO2 and subsequent catalytic turnover (Scheme
1). Additionally, if a strong acid is required, the catalyst should
operate at low potentials to limit background H+ reduction at
the working electrode, which can occur at the electrode surface
even in the absence of electrocatalyst at rates that increase
INTRODUCTION
■
There is currently great interest in developing more efficient
ways to use CO2 as a chemical feedstock for the production of
renewable fuels and value-added chemicals.1−3 Hydrocarbon
fuels are advantageous because of their ability to be stable, long-
term energy storage materials with higher energy densities than
batteries.4 The two-electron reduction of CO2 to produce CO
can be coupled with Fischer−Tropsch chemistry to produce
long-chain liquid hydrocarbon fuels, creating a closed CO2
cycle for combustion processes.4
CO2 + 2e− + 2H+ → CO + H2O
E° = −0.53 V at pH 7 vs NHE
2H+ + 2e− → H2
E° = −0.42 V at pH 7 vs NHE
The use of homogeneous molecular electrocatalysts to enable
the reduction of CO2 to CO is an attractive approach because
the properties of the electrocatalyst can be tuned through
ligand modification and well-defined complexes can be studied
in solution, allowing a greater understanding of the mechanism
of the reaction with CO2 to be attained. Research into CO2
reduction electrocatalysts has progressed since early discoveries
in the 1980s, although a robust understanding of the design
factors required for highly selective CO2 reduction over H+
reduction, an inherently competitive reaction because of the
requirement of protons in the reduction of CO2, is still
developing.3,5,6 Previous investigations probing the structure−
activity relationships of phosphine-containing palladium pincer
electrocatalysts found that selectivity for CO2 reduction was
extremely sensitive to variations in the ligand substituents,
Received: July 28, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.5b01698
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
with 1-iodobutane and then the addition of 2 equiv to 2,6-
bis(bromomethyl)pyridine, yielding the dibromide salt. Coor-
dination to palladium was achieved via transmetalation from a
silver carbene intermediate, where the proligand was reacted
with 1 equiv of silver(I) oxide in the presence of 3 Å molecular
sieves in DMSO and then with equivalents of PdCl2(cod) and
silver triflate once silver(I) oxide had reacted, yielding the
[Pd(L)Cl]OTf complexes 1 and 2 in 46−66% isolated yields
(Scheme 2). Purification can be achieved either by repeated
washes with small aliquots of methanol or by column
chromatography using a N,N-dimethylformimide (DMF)/
CH3CN/KNO3(aq) solvent mixture as the eluent. Once
synthesized, these complexes were functionalized by halide
abstraction using an excess of silver triflate in the presence of a
coordinating neutral ligand, in this case acetonitrile, yielding 1-
CH3CN and 2-CH3CN in 66−80% yields.
Scheme 1. Proposed Equilibria for Protonation of Activated
CO2
exponentially with applied overpotential when mass transport is
not limiting.12
Redox-active ligands have a promising role in the design of
improved CO2 reduction electrocatalysts because they can store
at least some of the charge required for CO2 activation away
from the metal center, reducing the basicity of the reduced
complex. This is showcased prominently in M(bpy)(CO)3X-
type complexes (where M = Mn or Re)13,14 and is an important
factor in previous work from our group where lutidyl-linked
bis(N-heterocyclic carbene) Pd pincer complexes were found
to reduce CO2 to CO at moderate potentials with trifluoro-
acetic acid [TFA; pKa = 3.5 in dimethyl sulfoxide (DMSO)15]
as the proton source.10 One mitigating factor for the
performance of these bis-NHC electrocatalysts is the direct
reduction of H+ at the electrode surface, accounting for up to
half of the current passed at typical operating potentials.
With this general bis-NHC ligand framework, less negative
reduction potentials and improved selectivity for CO2 over H+
reduction were observed for benzimidazol-2-ylidene relative to
imidazol-2-ylidene NHC donors.10 Modification of the
annulating moiety has been shown to considerably affect the
σ-donating and π-accepting properties of the NHC donor, both
of which are significant factors when bonded to a late-transition
metal,16,17 providing a means of tuning the electronic properties
of the complex. In this vein, we have synthesized Pd pincer
complexes with further extended π systems attached to the
NHC moiety, namely, phenanthro- and pyreno-annulated
NHCs, to investigate their relative performance as CO2
reduction electrocatalysts.17,18 Here we demonstrate that
these complexes are capable of reducing CO2 to CO and
have additional redox activity, which enables increased electron
donation and activation of CO2.
The complexes were characterized by mass spectrometry
(MS), 1H NMR, and attenuated total reflectance Fourier
transform infrared (ATR-FTIR) spectroscopy, with the IR
spectra indicating the presence of the bound acetonitrilo ligand
for 1-CH3CN and 2-CH3CN (Figures S7−S12). Microanalysis
of 2 yielded inconsistent results. Over four analyses, % N and %
H were consistent and within the acceptable range, but % C
was consistently 2.7−3.0% below calculated amounts. Micro-
analysis of the derivative product 2-CH3CN matched the
predicted % N, % H, and % S but was also 2.6−2.7% below
calculated amounts for % C, which cannot be accounted for by
the presence of unreacted starting materials or silver carbene
intermediates. Incomplete combustion of electropositive
complexes with M−C bonds has been observed previously
and attributed to carbide formation.19−21
RESULTS AND DISCUSSION
■
Synthesis and Structure. The proligands were synthesized
by condensation of phenanthrene-9,10-dione or pyrene-4,5-
dione with formaldehyde and ammonium acetate in acetic acid
to yield the polyannulated imidazole, followed by alkylation
Scheme 2. Synthetic Route to 1, 2, 1-CH3CN, and 2-CH3CN
B
DOI: 10.1021/acs.inorgchem.5b01698
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Both 1 and 2 have low solubility in many organic solvents
other than DMF and DMSO, presumably because of
intermolecular π-stacking interactions. The acetonitrilo com-
plexes 1-CH3CN and 2-CH3CN have improved, but still
limited, solubility in solvents such as acetonitrile, acetone, and
dichloromethane. Additionally, both proligands are highly
fluorescent under UV irradiation, whereas the palladium
complexes are nonemissive, providing a convenient means to
detect impurities. The absorption spectrum of 2 shows some
absorption at the red end of the spectrum, which could be
relevant for photocatalytic applications (Figure S13).
X-ray-quality crystals of 1 were grown by the slow diffusion
of diethyl ether into a saturated DMF/CH3CN solution.
Interestingly, syntheses of 1 with [BF4]− as the counterion
instead of [OTf]− resulted in the growth of modulated crystals,
which contained a periodically expanding and contracting unit
cell, requiring an unusual and complicated solution and
refinement process.22 Despite repeated attempts under a
variety of conditions, X-ray-quality crystals of 2 could not be
grown with either tetrafluoroborate or triflate counterions; fine
powders were obtained in all cases.
unit cell, the phenanthrene moieties of neighboring molecules
are aligned with an interplanar distance of 3.4−3.8 Å, which is
consistent with the presence of a weak π-stacking interaction.
Electrochemical Characterization under N2 and CO2.
Complexes 1 and 2 were electrochemically characterized by
cyclic and square-wave voltammetry under both N2 and CO2
atmospheres, revealing two prominent irreversible reduction
waves followed by a larger reduction feature that displays some
reversibility for 1 at scan rates above 1 V/s and is reversible for
2. The irreversible waves were tested for reversibility by
immediate postreduction oxidation, but none was detected.
The peak potentials for these waves were determined by
square-wave voltammetry and are reported in Table 1 and
a
Table 1. Peak Potentials versus Ferrocene0/+ for 1 and 2
complex
Epc1 (V)
Epc2 (V)
−2.18
−2.20
Epc3 (V)
−2.49
−2.30
Epc4 (V)
Epc5 (V)
−2.96
b
1
2
−1.77
−1.74
−2.83
−2.56 (rev)
b
b
a
Recorded from square-wave voltammograms at 25 Hz for 2 mM
solutions in 0.1 M [n-Bu4N]PF6/DMF under N2 using a glassy carbon
b
working electrode. Peak potentials determined by deconvolution.
The solid-state structure of 1 (Figure 1) revealed a Pd−
pyridyl bond length of 2.048 Å, a Pd−Cl bond length of 2.295
Figure 1. Solid-state structure of 1. The triflate anion and acetonitrile
within the unit cell are omitted.
Figure 2. Overlaid square-wave voltammograms of [Pd(C^N^C)Cl]-
BF4, [Pd(bC^N^bC)Cl]BF4, 1, and 2 at 25 Hz.
Å, an average Pd−NHC bond length of 2.038 Å, and a bite
angle of 172.0° for the pincer ligand, all of which are
comparable to similar complexes with other NHC groups,
although the Pd−pyridyl length is approximately 0.02 Å shorter
than that in analogous imidazol-2-ylidene and benzimidazol-2-
ylidene complexes.23−29 This may be due to increased π-back-
bonding into the π-acidic pyridyl moiety because the
phenanthro-annulated NHC has been shown to have increased
π charge density at the carbene center compared with benz-
annulated imidazol-2-ylidenes.17 In general, there is an
increasing degree of aromaticity in the imidazol-2-ylidene
moiety as the size of the annulated π-system increases, as seen
by a downfield shift for the adjacent methylenic protons in the
1H NMR spectra due to increased ring current, from 5.64 ppm
for the bridging methylenic protons in the nonannulated
species to 6.04, 6.63, and 6.80 ppm for benz-, phenanthro-, and
pyreno-annulated species, respectively.
Figure 2. The first two reduction events were found to occur at
potentials very similar to the reduction of [Pd(bC^N^bC)Cl]-
BF4, indicating that the trend of lower reduction potentials with
increasing size of the annulated π system did not extend
beyond [Pd(bC^N^bC)Cl]BF4. The incorporation of phenan-
threne and pyrene moieties does, however, lead to new
reduction features at approximately −2.92 and −2.57 V for 1
and 2, respectively, which can be assigned as one-electron
reductions of the two polyaromatic moieties present in each
complex because the peak current is twice that of the first
reduction event in both cases (Figure 3).
At the peak potential of the first reduction event in 1, a
smaller current enhancement in the presence of CO2 is
observed, compared to the current enhancements with
[Pd(C^N^C)Cl]BF4 and [Pd(bC^N^bC)Cl]BF4, and 2
showed no current enhancement at all. This may be due to
greater stabilization of the additional charge due to the
presence of extended π systems. A large current enhancement
was observed for each complex upon transfer of the second
The NHC moieties are twisted out-of-plane by approx-
imately 45°, comparable to the value in similar lutidine-linked
bis-NHC Pd complexes, limiting π stacking in solution and
enabling solubility in a few polar aprotic solvents. Within the
C
DOI: 10.1021/acs.inorgchem.5b01698
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 3. Cyclic voltammograms of 1 (left) and 2 (right) at 100 mV/s under N2 and CO2.
Figure 4. Cyclic voltammograms of 1-CH3CN (left) and 2-CH3CN (right) at 100 mV/s under N2 and CO2.
electron, indicative of chemical reactivity between the reduced
complex and CO2. A further current enhancement was
observed upon reduction of the polyaromatic NHC groups,
as well as a loss of electrochemical reversibility. This behavior is
consistent with an increased degree of CO2 activation due to
increased electron donation from the reduced NHC groups.
In addition, because of the size of the π systems in these
molecules, especially 2, it is conceivable that their electro-
chemical response could deviate from normal solution behavior
because of π interactions between the graphitic surface of the
glassy carbon electrode and the complexes. Thus, electro-
chemical experiments were also performed with a platinum
electrode, and these revealed the same characteristics and broad
waves as those with glassy carbon. In addition, with either
platinum or carbon electrodes and for both complexes, the peak
current increased linearly with the square root of the scan rate,
indicating freely diffusing species in solution according to the
Randles−Sevcik equation.12 Thus, the relative broadness of the
reduction waves, especially evident with 2, is attributed solely to
slower diffusion rates due to the increased size of the molecules
as well as some aggregation in solution.
imidazol-2-ylidenes. These experiments gave diffusion rates of
1.6 × 10−6 cm2/s for [Pd(bC^N^bC)Cl]BF4, 1.3 × 10−6 cm2/s
for 1, and 1.1 × 10−6 cm2/s for 2 in DMSO-d6 (Figures S15 and
S16). By a comparison of the conservative estimates of the
hydrodynamic radii based on the geometry-optimized DFT
structures, decreases in the diffusion rate of 18% from
[Pd(bC^N^bC)Cl]BF4 to 1 and then 8% from 1 to 2 are
predicted according to the Stokes−Einstein equation.12
Decreases of 20% and then 14%, respectively, are observed,
with this unaccounted-for decrease attributable to π-aggrega-
tion interactions, especially in the case of 2. This is also
consistent with the broadness of 1H NMR resonances for 1 and
especially 2, most prominent at high concentrations, in the
high-viscosity NMR solvent DMSO-d6.
The electrochemistry of the dicationic solvento complexes 1-
CH3CN and 2-CH3CN was also analyzed (Figure 4), revealing
in both cases an anodic shift in the peak potential of the first
cathodic wave, by +0.34 and +0.36 V, respectively, relative to
the chloro complexes 1 and 2 (Table 2). As observed
previously with related complexes,10 the dicationic species
were able to scavenge trace chloride ions in solution to replace
the labile solvento ligand. The addition of tetrabutylammonium
chloride to a solution of 1-CH3CN led to loss of the cathodic
wave at −1.43 V and growth at −1.77 V, matching the
electrochemical features of 1. The addition of tetraoctylammo-
nium bromide to a solution of 1-CH3CN led to loss of the
cathodic wave at −1.43 V and growth of a new wave at −1.61
1
To provide further support for this, H NMR diffusion-
ordered spectroscopy (DOSY) experiments were performed to
compare the relative diffusion rates of [Pd(bC^N^bC)Cl]BF4,
1, and 2 in DMSO-d6 because each of these complexes has
characteristic, nonoverlapping resonances for the diastereotopic
protons of the methylene groups linking the pyridyl rings to the
D
DOI: 10.1021/acs.inorgchem.5b01698
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 2. Peak Potentials versus Ferrocene0/+ of the First
Cathodic Wave for 1-CH3CN and 2-CH3CN Compared to
a
Those of 1 and 2
complex
Epc1 (V)
Δ(Epc1) (V)
1
−1.77
−1.43
−1.76
−1.40
+0.34
1-CH3CN
2
+0.36
2-CH3CN
a
Recorded from square-wave voltammograms at 25 Hz for 2 mM
solutions in 0.1 M [n-Bu4N]PF6/DMF under N2 using a glassy carbon
working electrode.
V, attributed to a bromo species. The reduction waves at −2.18
V and between −2.83 and −2.96 V are unchanged relative to 1
for both 1-CH3CN and 1-Br, and the wave at −2.49 V is no
longer present (or cathodically shifted). Thus, the potential of
the first electron transfer to the complex as well as the third
reduction peak in 1 is found to be sensitive to the ligand in the
fourth coordination site.
Figure 5. Overlaid cyclic voltammograms of 1 at 100 mV/s under N2
and CO2 with and without 25 mM Mg(ClO4)2 present.
Electrochemical Effects of Stabilizing Cations. Electro-
chemical characterization of 1 was also carried out in the
presence of Mg2+ through the addition of 25 mM Mg(ClO4)2.
Previous investigations have shown that the presence of alkali-
and alkali-earth-metal cations can assist in CO2 reduction by
stabilizing the charge buildup on the oxygen atoms of partially
activated CO2.10,30−32 In addition to this general CO2-
stabilizing effect, the presence of coordinating cations may be
particularly beneficial for electrocatalysts with a pincer motif,
where an initial loss of the ligand in the fourth coordination
site, often a halide, upon reduction is required for catalytic
turnover. For example, it has been shown that palladium
triphosphine pincer complexes require a labile solvento ligand
in the fourth coordination site for CO2 reduction activity.
Similarly, although there is CO2 reduction activity with a halide
ligand present for lutidine-linked bis-NHC Pd pincer
complexes, the performance is improved when the halide is
replaced by a solvento ligand.7,10 Thus, the presence of more
strongly interacting cations in solution may have a dual
function by also helping to abstract and/or passivate
coordinating anions. Both of these effects can be inferred
from a comparison of the cyclic voltammograms of 1 taken
under N2 and CO2 and with/without Mg2+ ions present (Figure
5). Without any CO2 in solution, the presence of Mg2+ ions led
to increased current at the second reduction wave, a peak that
likely results from the reduction of an electrogenerated species
with the chloro ligand dissociated. This assignment is also
supported by cyclic voltammograms taken at high scan rates
(1−5 V/s), where the second reduction wave diminishes as the
scan rate increases (attributed to a chloro-dissociated species),
and the third wave at −2.49 V becomes more prominent
(attributed to a chloro-bound species). With CO2 but no
proton source in solution, an attenuated current enhancement
is observed with Mg2+, which can be explained by its stabilizing
ability. With a proton source present, the activated CO2 can be
protonated and reduced.
and activation of CO2 as it approaches a reduced polyaromatic
complex.
A molecular orbital (MO) analysis of DFT-1 and DFT-2
shows a smaller highest occupied molecular orbital (HOMO)/
lowest unoccupied molecular orbital (LUMO) gap
(ΔHOMO−LUMO = 6.96 and 6.45 eV, respectively) for the
polyaromatic species compared to those of [Pd(C^N^C)Cl]+
(ΔHOMO−LUMO = 8.58 eV) and [Pd(bC^N^bC)Cl]+
(ΔHOMO−LUMO = 8.42 eV). There are also additional lower-
lying unoccupied MOs available in DFT-1 and DFT-2.
Additionally, the polyannulated NHCs were found to be
stronger electron donors than analogous (benz)imidazol-2-
ylidenes, yielding atomic charges of +0.176 and +0.175 on
palladium for DFT-1 and DFT-2 compared to +0.234 and
+0.250 for [Pd(C^N^C)Cl]+ and [Pd(bC^N^bC)Cl]+, re-
spectively, based on natural bond orbital (NBO) analyses. In
both model systems, the LUMO is localized on the lutidine
linker and contains some Pd dxz character (Figure 6). At higher
energies than that of the LUMO (approximately +0.75 eV) are
a cluster of distinct MOs at very similar energies, providing
information about where added electrons may reside on the
reduced complexes. Within this cluster, there is an orbital with
2
2
primarily Pd dx −y character and two orbitals with primarily
phenanthrene/pyrene character (Figure 6), showing that the
polyaromatic NHC ligands can potentially act as electron
reservoirs, providing additional redox activity to the pincer
proligand. Each of these orbitals are localized in distinct and
complementary areas of the molecule, providing potentially
good charge diffusion and stabilization when populated.
The relative energies of these orbitals and the order of their
population as electrons are added to the system will be
perturbed as additional charge is added, especially when there
are accompanying geometry changes in the complex. Thus,
geometry-optimized structures were calculated as additional
electrons were added to the system.
DFT Modeling and Investigations. To gain greater
insight into these complexes, DFT calculations were carried out
to (1) give optimized solution-state structures of the unreduced
and one-, two-, and three-electron-reduced species, (2)
determine an approximate degree of activation for CO2
bound to the reduced species, and (3) calculate the energetics
E
DOI: 10.1021/acs.inorgchem.5b01698
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 6. Calculated geometries of low-lying unoccupied MOs for DFT-2. LUMO+2 is similar in energy and geometry to LUMO+1 but also
contains some pyridyl character.
Figure 7. HOMO and SOMO geometries of lowest-energy geometry-optimized structures of reduced species of DFT-1 and DFT-2 as pristine,
nonhalide-dissociated species.
2.41 Å. At this point, the chloride has only a small energy
barrier to dissociation, which is calculated to be exergonic by
approximately 50 kcal/mol. Dissociation of the halide lowers
2
2
the energy of the empty dx −y orbital below that of the
polyaromatic moieties.
The third electron added to the system for both complexes,
even for halide-dissociated species, is localized on the
polyaromatic moieties. The LUMO is also phenanthro- or
pyreno-localized, which is consistent with the experimental
results showing multielectron transfer at the most negative
reduction potentials. The characteristics of the modeled
reduced species are consistent with the electrochemical data
and provide additional insight, suggesting stability of the Pd−
NHC bonds even in highly reduced species, weakening of the
halide bond by the first two pyridyl/palladium-centered
electron transfers, and additional redox activity on the
phenanthrene and pyrene moieties upon transfer of three and
four electrons.
Another redox-active annulated NHC ligand, naphthoquino-
noimidazol-2-ylidene, has been previously reported, showing
reversible reduction/oxidation in Ir(L)(CO)2Cl at −1.08 V, a
much less negative potential than pyrene (−2.54 V) or
phenanthrene (−2.82 V),33−35 where reduction of the redox-
active group switched the NHC to a more electron-donating
state.36 A similar effect would be expected upon reduction of
the polyaromatic groups fused to the NHC, providing
additional electron donation to a bound CO2 molecule. For
this reason, calculations of the degree of CO2 activation as a
function of the degree of reduction of the complex were
investigated and compared to those of other known electro-
catalysts. The bond angle of the bound CO2 at the energy
minimum of the reaction coordinate was used as a convenient
quantity to represent this, although analysis of the bonding
energy along the whole reaction coordinate is required for strict
comparisons. These bond angles are dependent on the chosen
computational parameters such as the functional and basis set
The first electron added to each of DFT-1 and DFT-2 was
found to be localized on the pyridyl group, with some Pd and
NHC character, and resulted in contraction of the Pd−N bond
(from approximately 2.05 to 2.01 Å) and lengthening of the
Pd−Cl bond (from approximately 2.30 to 2.38 Å), as well as
the lengthening and contraction of bonds within the pyridyl
group toward more distinct single bond/double bond character,
intermediate to dihydropyridine (Figure 7). The weakening of
the Pd−Cl bond is pertinent for catalysis, where displacement
of the ligand is necessary. Calculations of the free energy of
chloro dissociation were performed on DFT-1, DFT-2, and a
model complex of [Pd(C^N^CR=Me)Cl]+, yielding similar
results of +42 1 kcal/mol for the nonreduced species and
−1
1 kcal/mol for the one-electron-reduced species. The
dissociation energies were found to correlate with the basicity
of the ligand, where bromo, trifluoroacetate, and trifluoroetha-
nolate have free energies of dissociation of −4, −11, and +8
kcal/mol, respectively, from the one-electron-reduced species
of [Pd(C^N^CR=Me)X]+.
The second electron can be added in either a spin-aligned
(singlet) or antialigned (triplet) manner. Both configurations
were tested and found to be almost identical in energy, with the
singlet state being marginally more stable for DFT-1 and the
triplet state slightly more stable for DFT-2 (by 1−3 kcal/mol).
In the triplet state, the electron is localized on the polyaromatic
moieties, with the singly occupied molecular orbital (SOMO)
in DFT-2 being 0.42 eV lower in energy because of more
extensive π conjugation. In the singlet state, the additional
electron is localized across the palladium, NHC, and pyridyl
groups, resulting in a further lengthening of the Pd−Cl bond to
F
DOI: 10.1021/acs.inorgchem.5b01698
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
and are best interpreted as relative assessments of CO2
activation.
The results of these calculations show an increase in the
degree of CO2 activation by species with reduced polyaromatic
moieties, from a bond angle of approximately 128° in two-
is localized on the pyridyl group, bound CO2, and metal center;
in the three-electron-reduced case, most of the additional
charge is located on the periphery of the polyaromatic groups;
in the four-electron-reduced case, most of the additional charge
is localized on the pyridyl group and metal center, as well.
Last, the thermodynamics of CO2 coordination to one- and
two-electron-reduced DFT-1 was modeled as a function of the
distance from the metal center. Compared to [Pd(C^N^C)-
Cl]+, the singly reduced species has a slightly more exergonic
Table 3. Calculated Degree of CO2 Activation When Bound
to Reduced Complexes
complex
degree of reduction
CO2 bond angle (deg)
DFT-1
2e−
3e−
4e−
2e−
3e−
4e−
127.6
125.8
123.8
128.2
127.9
123.9
DFT-2
electron-reduced species to 124° in four-electron-reduced
species (Table 3). A comparison of the degree of CO2
activation as determined by the bond angle, when modeled as
bound to an active electrocatalyst species, to the proton sources
used in CO2 reduction with the particular electrocatalyst shows
a correlation between the degree of activation and pKa of the
proton source required; highly activated CO2 can be
Table 4. Correlation between the Calculated Degree of CO2
Activation at the Bonding Potential Energy Minimum and
Required Proton Source6,7,10,13
Figure 8. DFT-calculated energetics for the activation of CO2 as it
approaches a reduced metal complex. Bond angles of CO2 are given at
the energy minimum, unless otherwise stated.
reduced complex
CO2 bond angle (deg)
proton source
−
Re(bpy)(CO)3
122.5
123.7
130.0
128.0
adventitious H’s, H2O
TFE, methanol, water
HBF4, H3PO4
−
Mn(bpy)(CO)3
interaction, and the doubly reduced species is almost equivalent
(Figure 8). Because of the computational expense, only the
one- and two- electron-reduced species of DFT-1 were
modeled in this way, with only the optimized structures at
the energy minima calculated for three- and four-electron-
reduced DFT-1 and DFT-2 (see above).
Controlled Potential Electrolysis (CPE). To survey the
relative performance of the complexes, CPE experiments were
performed at a variety of potentials with CO2-sparged 2 mM
solutions in 0.1 M [n-Bu4N]PF6/DMF with 10 mM TFA
present, unless otherwise specified. The headspace gas was then
sampled and analyzed by gas chromatography to quantify the
CO and H2 produced.
Pd(triphos)0
Pd(C^N^C)0
TFA
protonated by weak acids, and weak acids lead to less undesired
H+ reduction in CO2 electroreduction experiments (Table 4).
For example, Mn(bpy)(CO)3(X) requires the use of weak acids
like 2,2,2-trifluoroethanol (TFE; pKa = 23.5 in DMSO15) to
reduce CO2, whereas reduction using Re(bpy)(CO)3(X) can
proceed without an added proton source because of the
difference in the CO2 activation capability, calculated to bond
angles of 123.7° and 122.5° at their potential energy minima,
respectively. Thus, it is reasonable that four-electron-reduced
species of 1 and 2 may be able to reduce CO2 in the presence
of a weak acid, such as TFE. Accordingly, cyclic voltammo-
grams of 2 and 2-CH3CN were taken under CO2, with the
addition of TFE resulting in increased current at the wave
attributed to reduction of the pyrene moieties (Figure S17),
consistent with these theoretical results.
The optimized structures for the highly reduced CO2-bound
species of DFT-1 and DFT-2 were chemically reasonable with
square-planar geometry, conventional Pd−L bond lengths, and
CO2 occupying the fourth coordination site. Conversely,
optimizations of simpler cationic complexes such as [Pd-
(C^N^C)]+ and [Pd(bC^N^bC)]+ with three and four
electrons added result in distortions toward a tetrahedral
geometry, loss of planarity in the pyridyl moiety, and loss of
symmetry in pincer coordination. An NBO charge analysis of
the optimized structures revealed increased electron density on
the NHC moieties relative to the unreduced complex for the
three- and four-electron-reduced cases, indicating charge
storage. In the four-electron-reduced case, the additional charge
Electrolysis at the first reduction potential of 1 led to 18%
CO in the gas produced (Table 5). This ratio decreased to 12%
at the second reduction potential and then increased to 20% at
a
Table 5. CPE Results for 1
potential
(V)
charge
passed (C)
ratio of CO
produced (%)
FE CO FE H2
solution
(%)
(%)
1
−2.02
−2.32
−2.97
−2.02
−2.32
−2.77
−2.02
3.7
3.7
3.8
3.0
3.7
4.1
68
18
12
20
27
14
16
28
17
10
15
26
11
14
27
79
68
61
70
65
74
69
1
1
1 + Mg2+
1 + Mg2+
1 + Mg2+
b
1
a
Performed with a 2 mM solution in 10 mL of 0.1 M [n-Bu4N]PF6/
DMF under CO2 using a glassy carbon rod working electrode.
Potentials versus ferrocene0/+. 25 mM Mg(ClO4)2 used, where
b
applicable. A concentration of 1.6 mM 1 was used.
G
DOI: 10.1021/acs.inorgchem.5b01698
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
more extreme potentials, where the phenanthrene moieties
could be reduced, even though the overpotential for H+
reduction at the electrode is very high. The addition of Mg2+
improved the selectivity for CO2 reduction during electrolysis
at the first reduction potential, but selectivity at more negative
potentials was either comparable or slightly worse.
formation of an insoluble species by chemical reaction with
TFE or its conjugate base.
Electrolysis with 1-CH3CN revealed an initial ratio of 25%
CO produced at −1.67 V, which is an improvement over 17%
with the analogous [Pd(C^N^C)(CH3CN)]2+ species.10 This
selectivity also seems to be short-lived, however, because the
characteristic electrochemical features for the solvento complex
were diminished after the first electrolysis experiment, which
may be due to coordination of the generated trifluoroacetate
anions. Subsequent experiments showed low selectivities for
CO2 reduction regardless of the potential applied and whether
or not Mg2+ was added (Table 7). This is in contrast to 1 and
also to the simpler complex [Pd(C^N^C)(CH3CN)]2+, which
A longer experiment was also performed where 68 C was
passed at −2.02 V and the headspace gas was monitored over
six intervals. The ratio of CO produced was consistently
between 29 and 32% over the first 56 C and 23% over the last
12 C, representing 5.5 turnovers of CO2 reduction to produce
CO. Cyclic voltammograms taken throughout the experiment
retain characteristic features of the complex (Figure S14).
Additional TFA was added at intervals of approximately 8 C to
maintain its concentration between 2 and 10 mM. The higher
efficiency for CO production observed in this longer experi-
ment is attributed to the smaller average concentration of acid
present as well as achievement of a steady-state concentration
of the active, reduced species. The shorter experiments serve to
screen the relative performance at different reduction
potentials.
a
Table 7. CPE Results for 1-CH3CN
charge
passed
(C)
ratio of CO
produced
(%)
FE
CO
(%)
FE
H2
potential
(V)
solution
1-CH3CN
(%)
−1.67
−1.97
−2.32
−1.67
−1.62
−1.97
−2.17
−2.32
3.5
3.5
3.5
3.1
3.5
3.8
3.8
3.8
25
14
5
24
12
4
72
78
75
94
90
87
85
80
b
1-CH3CN
1-CH3CN
1-CH3CN
1-CH3CN + Mg2+
1-CH3CN + Mg2+
1-CH3CN + Mg2+
1-CH3CN + Mg2+
5
5
7
7
a
Table 6. CPE Results for 2
7
7
6
6
charge
passed
(C)
ratio of CO
produced
(%)
FE
CO
(%)
FE
H2
(%)
potential
(V)
2
2
solution
a
Performed in sequence from the top to bottom row with a 2 mM
2
2
−1.97
−2.32
−2.02
−2.32
−2.62
−2.77
3.8
3.5
4.2
4.0
4.0
12
7
12
6
88
87
89
86
67
5
solution in 10 mL of 0.1 M [n-Bu4N]PF6/DMF under CO2 using a
glassy carbon rod working electrode. Potentials versus ferrocene0/+. 25
b
2 + Mg2+
11
12
13
87
11
12
10
33
mM Mg(ClO4)2 used, where applicable. The distinctive electro-
2 + Mg2+
2 + Mg2+
chemical feature for the acetonitrilo complex diminished after CPE at
−1.67 V.
b
2 + 200 mM TFE
1.5
a
maintained selectivity for CO2 throughout of 6 times the charge
transfer.10
Performed with a 2 mM solution in 10 mL of 0.1 M [n-Bu4N]PF6/
DMF under CO2 using a glassy carbon rod working electrode.
Potentials versus ferrocene0/+. 25 mM Mg(ClO4)2 used, where
Because of deactivation of 1-CH3CN during electrolysis with
TFA as the proton source and given that its first reduction peak
is at −1.43 V, with the current onset at approximately −1.31 V,
electrolysis was attempted with the strong acid HPF6 (pKa ∼ 0
in DMSO15,37) as the proton source, given that its conjugate
base is the noncoordinating anion [PF6]−. Electrolysis at −1.67
and −1.47 V yielded 6% and 5% CO production, respectively,
b
applicable. The current diminished very quickly as the film formed
on the electrode. A potential step to +0.7 V removes the film and
recovers a moderate current, which quickly diminishes at −2.77 V.
Electrolysis of 2 yielded similar results, with 12% CO
produced at the first reduction potential and a decrease to 7%
at the second reduction potential (Table 6). The addition of
Mg2+ did not result in the same increase in selectivity at the first
reduction potential observed with 1, although there was a small
improvement during electrolysis at the second and third
reduction potentials. Because of the more accessible potential
for reduction into the redox-active pyrenoimidazol-2-ylidenes
and the effect this could have on the degree of CO2 activation,
electrolysis was attempted at −2.77 V using 200 mM TFE as
the proton source. This resulted in selective but short-lived
CO2 reduction, where the current diminished within
approximately 15 s of electrolysis. After electrolysis, it was
observed that a dark film had deposited on the working
electrode. A potential step to +0.7 V removed the film and
recovered the initial current, which quickly diminished again
upon reduction at −2.77 V. This process was repeated several
times and the headspace gas then sampled, revealing 87% CO
in the gas produced, which is consistent with greater CO2
activation resulting from reduction of the high-energy redox-
active pyrene moieties. Given that the formation of this film
with resulting current deterioration is not observed at these
potentials using TFA as the acid source, it may be due to the
a
Table 8. CPE Results for 1-CH3CN with 10 mM HPF6
charge
passed
(C)
ratio of CO
produced
(%)
FE
CO
(%)
FE
H2
potential
(V)
solution
1-CH3CN
(%)
−1.67
−1.47
−2.32
−1.82
−2.32
4.1
3.0
2.2
3.0
3.0
6
5
3
4
4
6
5
2
4
3
89
91
73
97
83
1-CH3CN
1-CH3CN
1-CH3CN + Mg2+
1-CH3CN + Mg2+
a
Performed with a 2 mM solution in 10 mL 0.1 M [n-Bu4N]PF6/DMF
under CO2 using a glassy carbon rod working electrode. Potentials
versus ferrocene0/+. 25 mM Mg(ClO4)2 used, where applicable.
with H+ reduction predominant with this proton source (Table
8). Despite the poor selectivity for CO2 reduction, the cyclic
voltammograms of the complex before and after electrolysis
experiments were more consistent, suggesting that inhibition of
the complex with the conjugate base of the proton source may
be a relevant factor.
H
DOI: 10.1021/acs.inorgchem.5b01698
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Finally, separate electrolysis experiments were carried out
with 2-CH3CN using TFE and TFA as proton sources (Table
affect the reactivity. The solvento complexes 1-CH3CN and 2-
CH3CN were found to be more quickly deactivated during
electrolysis. The addition of polyaromatic groups enabled
additional redox activity on the ligand at −2.96 V for 1 and
−2.56 V for 2, allowing for increased electron donation from
the ligand. The addition of the weak acid TFE results in
reactivity with CO2 at these potentials, but CPEs are short-lived
and result in the formation of an insulating and insoluble
species on the electrode. These systems were also modeling
computationally, providing support for assignments of the
redox events, giving a greater understanding of the chemical
behavior of the reduced species, and highlighting the
importance of anion dissociation energies within the pincer
motif. Investigations into the effects of modifying the pyridyl
moiety and metal center are currently underway.
a
Table 9. CPE Results for 2-CH3CN
charge
passed
(C)
ratio of CO
produced
(%)
FE
CO
(%)
FE
H2
(%)
potential
(V)
solution
b
2-CH3CN + TFE
2-CH3CN + TFE
2-CH3CN
−2.52
−2.72
−2.07
−2.47
−2.07
−2.47
−2.72
0.5
100
100
4
12
14
4
0
0
b
1.0
2.6
2.5
3.1
3.2
3.2
84
73
74
59
49
2-CH3CN
6
4
2-CH3CN + Mg2+
2-CH3CN + Mg2+
2-CH3CN + Mg2+
2
1
12
9
8
5
a
Performed with a 2 mM solution in 10 mL of 0.1 M [n-Bu4N]PF6/
DMF under CO2 using a glassy carbon rod working electrode.
Potentials versus ferrocene0/+. 500 mM TFE and 25 mM Mg(ClO4)2
EXPERIMENTAL SECTION
■
b
used, where applicable. The current diminished quickly as the film
General Procedures. Unless otherwise specified, all reactions
were performed under nitrogen using standard Schlenk techniques and
solvents and reagents were used as received from commercial sources.
Magnesium perchlorate (Alfa), 2,2,2-trifluoroethanol (TFE; Aldrich),
and trifluoroacetic acid (TFA; Aldrich) were used as received for
electrochemical experiments. N,N-Dimethylformamide (DMF; EMD
Millipore Omnisolv) and distilled acetonitrile (Anachemia Accusolv)
were dried and stored over 15% (m/v) 4 Å molecular sieves.40,41
1H NMR spectra were acquired using Bruker AV300 or AV400
spectrometers with chemical shifts referenced to residual solvent
signals. Mass spectra were acquired using a Waters LC-MS
electrospray ionization mass spectrometer, except for the acetonitrilo
complexes, which were acquired using a Waters Micromass LCT
electrospray ionization mass spectrometer. IR spectra were collected
using a PerkinElmer Frontier FTIR spectrometer with an ATR
attachment. Gaseous products were analyzed using an SRI model
8610C gas chromatograph equipped with molecular-sieve columns,
dual thermal conductivity and flame ionization detectors, and a
methanizer. Crystallographic data were acquired using a Bruker X8
APEX II diffractometer with graphite-monochromated Mo Kα
radiation.
Electrochemistry. Electrochemical experiments were performed
using a Metrohm Autolab PGSTAT12 or Pine AFCBP1 bipotentiostat
in an airtight three-electrode cell with a 7 mm2 glassy carbon working
electrode (Bioanalytical Systems, Inc.), platinum mesh counter
electrode, and silver wire pseudoreference electrode in a 0.010 M
AgNO3 acetonitrile solution separated from the bulk solution by a
Vycor frit. Experiments were performed under N2 or CO2 using 2 mM
concentrations of the complexes in 10 mL of an anhydrous electrolyte
solution unless otherwise stated. CPE experiments used glassy carbon
rod (Alfa Aesar) working electrodes in a two compartment H cell,
where the counter electrode compartment was separated from the
compartment containing the working and reference electrodes by
fritted glass. All glassy carbon electrodes were cleaned by successive
polishing with 1, 0.3, and 0.05 μM alumina paste, followed by rinsing
with water, sonication (5 min) in distilled water, and sonication (5
min) in methanol. Electrolyte solutions were 0.10 M triply
recrystallized [n-Bu4N]PF6 in anhydrous DMF and sparged with
nitrogen prior to use. For experiments with CO2, the solution was
sparged with CO2 for 15 min. Decamethylferrocene was added at the
end of the electrochemical experiments as an internal standard,
showing a reversible redox couple at −404 mV vs Ag/AgNO3 and
−476 mV vs ferrocene/ferrocenium in 0.1 M [n-Bu4N]PF6/DMF.42
Cyclic voltammograms were collected at a scan rate of 100 mV/s and
square-wave voltammograms at a frequency of 25 Hz unless otherwise
stated.
formed on the electrode. A potential step to +0.7 V removes the film
and recovers a moderate current, which quickly diminishes at these
potentials.
9). With TFE, behavior similar to that for 2 was observed, with
the formation of a film on the electrode surface and a quickly
diminishing current, resulting in the reduction of CO2 to CO.
Again, holding the working electrode at +0.7 V resulted in
removal of the film and a temporarily restored current. With
TFA, selectivities for CO production were lower than those
with 1, 2, or 1-CH3CN at 4−12%, and a noticeable amount of a
dark precipitate was formed. This insoluble black powder was
isolated by centrifugation, dried, and analyzed by ATR-FTIR,
revealing features similar to those of 2 but with a strong, broad
stretch centered at 1854 cm−1 (Figure S15), tentatively
assigned to a carbonate or bicarbonate species.
Colorimetric spot tests with chromotropic acid and
methylquinaldinium38,39 were performed on the solutions
after electrolysis to test for the presence of formic acid,
formate, or formaldehyde; none of these species were detected.
Additionally, two CPE experiments were performed with 2-
CH3CN in acetonitrile, with the resulting solution analyzed by
1H NMR in CD3CN with suppression of the acetonitrile
solvent signal (1.96 ppm), which is well separated from signals
for formate/formic acid (8.0 ppm) and formaldehyde (9.6
ppm). Again, none of these potential reduction products was
detected. Finally, two experiments were performed to see if
heterogeneous catalysis may be a significant contributor to the
production of CO: a mercury drop test was performed with 1 at
−2.05 V, revealing no difference in the performance, and after
an electrolysis experiment with 1, the working electrode was
transferred to a new electrolyte solution with 10 mM TFA but
no dissolved complex. The passage of 5 C resulted in a ratio of
only 1% CO produced.
CONCLUSION
■
Phenanthro- and pyreno-annulated bis-NHC Pd pincer
complexes 1 and 2 were synthesized and found to have
reduction potentials similar to that of [Pd(bC^N^bC)Cl]+.
These complexes are able to electrocatalytically reduce CO2 to
CO in the presence of TFA, with 1 having appreciably
improved selectivity for CO2 reduction at the first reduction
potential compared to [Pd(C^N^C)Cl]+ and [Pd(bC^N^bC)-
Cl]+ and 2 having diminished relative selectivity, showing that
electronic modifications to the NHC donors can drastically
Computational Methods. DFT calculations were performed
using Gaussian 09 (revision D.01) with the long-range and dispersion-
corrected ωB97xD hybrid functional without symmetry constraints.43
Calculations for reduced species were performed using unrestricted,
open-shell wave functions. The D95(d) basis set was used for all atoms
I
DOI: 10.1021/acs.inorgchem.5b01698
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
except transition metals, which employed the Stuttgart−Dresden−
Bonn quasi-relativistic effective-core potential and corresponding
correlation-consistent triple-ζ basis set.44,45 Calculations were
performed with the presence of a solvent reaction field of acetonitrile
produced by the conductor-like polarizable continuum model.46
Frequency calculations were performed on all geometry-optimized
structures to ensure that energy minima were achieved.
Synthesis. The proligand precursor 9,10-phenanthrenequinone
(95%) was purchased from Sigma-Aldrich, with the remaining
precursors synthesized according to literature procedures: 2,6-
bis(bromomethyl)pyridine,47 1H-phenanthro[9,10-d]imidazole,17 pyr-
ene-4,5-dione,48 and 9H-pyreno[4,5-d]imidazole.18
Synthesis of Halide Complexes. [Pd(phenC^N^phenC)Cl]OTf (1).
To a solution of L1 (0.369 mmol, 300 mg) in DMSO (12 mL) was
added Ag2O (0.368 mmol, 85.3 mg) and 3 Å molecular sieves. The
reaction vessel was covered in foil and left to stir at 50 °C for 24 h and
then left to cool to room temperature. To the resulting mixture was
added PdCl2(cod) (0.369 mmol, 105 mg) and then AgOTf (0.405
mmol, 104 mg). The mixture was left to stir for 48 h, then centrifuged,
and filtered through Celite. The solvent was removed under vacuum at
55 °C and the residue taken up in acetonitrile (5 mL). Addition to
Et2O (12 mL) yielded a light-yellow precipitate, which was isolated by
centrifugation and washed with Et2O (5 mL), MeOH (2 mL), and
Et2O (5 mL), yielding the crude product (260 mg). Purification was
achieved by successive washes with a minimum of MeOH (4 × 1 mL),
yielding a light-tan powder (160 mg, 46% yield). X-ray-quality crystals
were grown by the slow diffusion of Et2O into a concentrated 1:1
DMF/CH3CH solution at −30 °C. 1H NMR (400 MHz, (CD3)2SO):
δ 0.94 (t, J = 7.3 Hz, 6 H), 1.31−1.43 (m, 2 H), 1.55−1.68 (m, 2 H),
1.87−1.99 (m, 2 H), 2.00−2.12 (m, 2 H), 5.24−5.36 (m, 2 H), 5.48−
5.63 (m, 2 H), 6.41 (d, J = 15.9 Hz, 2 H), 6.90 (d, J = 15.4 Hz, 2 H),
7.79−7.95 (m, 8 H), 8.16 (s, 3 H), 8.57 (d, J = 8.2 Hz, 2 H), 8.99 (dt, J
= 8.4 and 0.7 Hz, 2 H), 9.07 (d, J = 8.2 Hz, 2 H), 9.08 (d, J = 8.2 Hz, 2
H). ESI-MS: m/z 792.6 ([M − OTf]+). Anal. Calcd for
C46H41ClF3N5O3PdS: C, 58.60; H, 4.38; N, 7.43. Found: C, 58.09;
H, 4.41; N, 7.19. Averaged results for repeat elemental analysis are
shown. For one attempt, the C was within error of the calculated value,
but for the other, it was below. This is accounted for by the presence
of a small amount (∼7%) of bromo species present, as shown by ESI-
MS at m/z 838 ([MBr − OTf]+), as well as some intractable DMSO (3
mol % by NMR), giving calculated values of 58.36% C, 4.38% H, and
7.39% N.
Alkylation of Polyaromatic Imidazoles. 1-Butyl-1H-phenanthro-
[9,10-d]imidazole. The compound was synthesized via a modified
literature procedure.17 Sodium hydroxide (4.40 mmol, 176 mg) was
added to a 10 mL solution of 1H-phenanthro[9,10-d]imidazole (3.99
mmol, 870 mg) in DMSO and stirred for 2 h at room temperature,
after which 1-iodobutane (4.5 mmol, 0.51 mL) was added and the
solution was heated to 50 °C overnight. Water (15 mL) was added to
the solution and extracted with diethyl ether (15 mL × 4). The organic
extracts were washed with water (15 mL × 2) and then the solvent was
removed, yielding a yellow-orange oil (899 mg). The crude product
was purified by column chromatography over SiO2 using acetone as
the eluent, collecting the second band. Upon removal of the solvent,
1
the oil crystallized as a yellow-orange solid (813 mg, 74% yield). H
NMR (300 MHz, (CD3)2CO): δ 0.98 (t, J = 7.4 Hz, 3 H), 1.48 (sxt, J
= 7.5 Hz, 2 H), 2.02 (quin, J = 7.5 Hz, 2 H), 4.78 (t, J = 7.2 Hz, 2 H),
7.55−7.79 (m, 4 H), 8.12 (s, 1 H), 8.41 (dt, J = 8.0 and 0.9 Hz, 1 H),
8.62−8.71 (m, 1 H), 8.81 (dt, J = 8.2 and 0.7 Hz, 1 H), 8.93 (dt, J =
8.2 and 0.8 Hz, 1 H).
9-Butyl-9H-pyreno[4,5-d]imidazole. The compound was synthe-
sized via a modified literature procedure.18 Sodium hydroxide (3.90
mmol, 156 mg) was added to a 12 mL solution of 9H-pyreno[4,5-
d]imidazole (3.55 mmol, 860 mg) in DMSO and stirred for 2 h at
room temperature, after which 1-iodobutane (4.0 mmol, 0.45 mL) was
added and the solution was heated to 50 °C overnight. Water (15 mL)
was added to the solution and extracted with diethyl ether (25 mL ×
5). The organic extracts were washed with water (20 mL × 2) and
dried with MgSO4, and the solvent was then removed, yielding an
orange crystalline powder (589 mg, 56% yield), which was used
[Pd(pyreC^N^pyreC)Cl]OTf (2). To a solution of L2 (0.319 mmol,
275 mg) in DMSO (12 mL) was added Ag2O (0.324 mmol, 75.1 mg)
and 3 Å molecular sieves. The reaction vessel was covered in foil, left
to stir at 55 °C for 24 h, and then left to cool to room temperature. To
the resulting mixture was added PdCl2(cod) (0.319 mmol, 99.1 mg)
and then AgOTf (0.351 mmol, 90.2 mg). The mixture was left to stir
for 56 h, then centrifuged, and filtered through Celite. The solvent was
removed under vacuum at 60 °C and the dark residue taken up in
acetonitrile (5 mL). Addition to Et2O (10 mL) yielded a yellow-
orange precipitate, which was isolated by centrifugation and washed
with Et2O (5 mL), MeOH (4 × 2 mL), and Et2O (5 mL), yielding a
tan powder (215 mg, 68% yield). The solutes from methanol washes
1
without further purification. H NMR (300 MHz, (CD3)2SO): δ 0.89
(t, J = 7.3 Hz, 3 H), 1.38 (sxt, J = 7.3 Hz, 2 H), 1.93 (quin, J = 7.2 Hz,
2 H), 4.78 (t, J = 7.1 Hz, 2 H), 8.07−8.14 (m, 2 H), 8.17 (s, 2 H),
8.19−8.27 (m, 2 H), 8.36 (s, 1 H), 8.57 (d, J = 7.6 Hz, 1 H), 8.80 (dt, J
= 7.5 and 0.6 Hz, 1 H).
1
were analyzed by NMR to track the degree of purification. H NMR
(400 MHz, (CD3)2SO): δ 0.99 (t, J = 7.4 Hz, 6 H), 1.39−1.55 (m, 2
H), 1.64−1.79 (m, 2 H), 2.09 (s, 2 H), 2.11−2.24 (m, 2 H), 5.40−5.53
(m, 2 H), 5.64−5.81 (m, 2 H), 6.53 (d, J = 15.9 Hz, 1 H), 7.12 (d, J =
15.7 Hz, 2 H), 8.18−8.38 (m, 11 H), 8.47 (d, J = 7.8 Hz, 2 H), 8.50
(d, J = 7.5 Hz, 3 H), 8.88 (d, J = 8.4 Hz, 2 H), 9.33 (d, J = 8.2 Hz, 2
H). ESI-MS: m/z 840.6 ([M − OTf]+). Anal. Calcd for
C50H41ClF3N5O3PdS: C, 60.61; H, 4.17; N, 7.07. Found: C, 57.76;
H, 4.06; N, 7.21.
Synthesis of Proligands. phenC^N^phenC·2HBr (L1). A solution
of 1-butyl-1H-phenanthro[9,10-d]imidazole (2.96 mmol, 813 mg) and
2,6-bis(bromomethyl)pyridine (1.40 mmol, 372 mg) in 20 mL of 1,4-
dioxane was heated to reflux for 84 h, during which time an off-white
precipitate was formed. The precipitate was collected by centrifuga-
tion, washed with diethyl ether (10 mL × 2), and then dried under
1
vacuum, yielding an off-white powder (732 mg, 64% yield). H NMR
Synthesis of Acetonitrilo Complexes. [Pd(phenC^N^phenC)-
(CH3CN)](OTf)2 (1-CH3CN). Complex 1 (0.050 mmol, 47 mg) was
taken up in a solvent mixture of 20 mL of 1:1 CH3CN/CH2Cl2, and
the light-yellow solution was reacted with AgOTf (0.11 mmol, 27 mg)
at room temperature overnight in the dark, forming a fine white
precipitate. The precipitate was removed by centrifugation and the
solvent removed under reduced pressure, leaving a light-yellow
residue, which was taken up in CH2Cl2 (3 mL), reprecipitated by
the addition of Et2O, washed with Et2O (2 × 3 mL), and then dried
(300 MHz, (CD3)2SO): δ 0.97 (t, J = 7.3 Hz, 6 H), 1.43 (sxt, J = 7.3
Hz, 4 H), 1.74 (quin, J = 7.4 Hz, 4 H), 4.38 (t, J = 7.4 Hz, 4 H), 6.14
(s, 4 H), 7.23 (t, J = 7.8 Hz, 2 H), 7.55 (t, J = 7.8 Hz, 2 H), 7.66 (d, J =
8.4 Hz, 2 H), 7.71−7.85 (m, 6 H), 8.05 (d, J = 8.1 Hz, 2 H), 8.20 (t, J
= 7.8 Hz, 1 H), 8.46 (d, J = 8.4 Hz, 2 H), 8.54 (d, J = 8.3 Hz, 2 H),
9.47 (s, 2 H). ESI-MS: m/z 732.8 ([M − Br]+).
pyreC^N^pyreC·2HBr (L2). A solution of 9-butyl-9H-pyreno[4,5-
d]imidazole (1.46 mmol, 436 mg) and 2,6-bis(bromomethyl)pyridine
(0.710 mmol, 188 mg) in 15 mL of 1,4-dioxane was heated to reflux
for 84 h, during which a light-colored precipitate was formed. The
precipitate was collected by centrifugation, washed with diethyl ether
(10 mL × 2), and then dried under vacuum, yielding an off-white
powder (411 mg, 67% yield). 1H NMR (300 MHz, (CD3)2SO): δ 0.98
(t, J = 7.3 Hz, 6 H), 1.44 (sxt, J = 7.4 Hz, 4 H), 1.71 (quin, J = 7.4 Hz,
4 H), 4.25 (t, J = 7.5 Hz, 4 H), 6.23 (s, 4 H), 7.54 (t, J = 7.9 Hz, 2 H),
7.80−8.04 (m, 14 H), 8.20 (dt, J = 7.3 and 0.8 Hz, 2 H), 8.29 (t, J =
7.8 Hz, 1 H), 9.58 (s, 2 H). ESI-MS: m/z 780.9 ([M − Br]+).
1
under vacuum, yielding an off-white powder (36 mg, 66% yield). H
NMR (400 MHz, CD3CN): δ 0.98 (t, J = 7.4 Hz, 6 H), 1.37−1.51 (m,
2 H), 1.59−1.73 (m, 2 H), 1.92−2.07 (m, 4 H), 1.96 (s, 3 H), 5.00
(dt, J = 14.9 and 8.1 Hz, 2 H), 5.23 (dt, J = 14.9 and 6.2 Hz, 2 H), 6.29
(d, J = 15.7 Hz, 2 H), 6.75 (d, J = 15.9 Hz, 2 H), 7.81−8.01 (m, 10 H),
8.09 (t, J = 8.0 Hz, 1 H), 8.54 (dt, J = 8.2 and 0.7 Hz, 2 H), 8.80 (dt, J
= 8.3 and 0.7 Hz, 2 H), 8.98 (d, J = 8.2 Hz, 2 H), 8.99 (d, J = 8.2 Hz, 2
H). ESI-MS: m/z 377.9 ([M − CH3CN − 2(OTf)]2+), 905.6 ([M −
CH3CN − OTf]+).
J
DOI: 10.1021/acs.inorgchem.5b01698
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
[Pd(pyreC^N^pyreC)(CH3CN)](OTf)2 (2-CH3CN). Complex 2
(0.082 mmol, 81 mg) was taken up in a solvent mixture of 20 mL
of 1:1:0.1 CH3CN/CH2Cl2/DMF. The dark-orange solution was
reacted with AgOTf (0.15 mmol, 38 mg) at room temperature for 4 h
in the dark, forming a light-yellow precipitate. The precipitate was
removed by centrifugation and the solution concentrated to ∼1 mL
under reduced pressure. Et2O (5 mL) was added, inducing
precipitation of a sticky orange solid, which was collected and
triturated with CHCl3 (2 mL) and washed with Et2O (2 × 3 mL). The
resulting light-orange powder was dissolved in CH3CN (2 mL) and
precipitated by the addition of Et2O (5 mL). The tan powder was
dried under a stream of N2 and then under vacuum, yielding a tan
powder (75 mg, 80% yield). 1H NMR (400 MHz, CD3CN): δ 1.05 (t,
J = 7.4 Hz, 6 H), 1.48−1.63 (m, 2 H), 1.73−1.85 (m, 2 H), 1.91−2.02
(m, 2 H), 1.96 (s, 3 H), 2.08−2.20 (m, 2 H), 5.10−5.22 (m, 2 H),
5.37−5.49 (m, 2 H), 6.45 (d, J = 15.9 Hz, 2 H), 6.99 (d, J = 15.7 Hz, 2
H), 8.08 (s, 4 H), 8.29−8.39 (m, 7 H), 8.47 (d, J = 7.7 Hz, 2 H), 8.50
(d, J = 7.8 Hz, 2 H), 8.87 (d, J = 8.0 Hz, 2 H), 9.16 (d, J = 8.0 Hz, 2
H). ESI-MS: m/z 402.7 ([M − CH3CN − 2(OTf)]2+). Anal. Calcd for
C52H43F6N6O6PdS2: C, 55.15; H, 3.83; N, 7.42; S, 5.60. Found: C,
52.49; H, 3.78; N, 7.40; S, 5.36.
(7) DuBois, D. L.; Miedaner, A.; Haltiwanger, R. C. J. Am. Chem. Soc.
1991, 113, 8753.
(8) Dubois, D. L. Comments Inorg. Chem. 1997, 19, 307.
(9) DuBois, D. L.; Miedaner, A. J. Am. Chem. Soc. 1987, 109, 113.
(10) Therrien, J. A.; Wolf, M. O.; Patrick, B. O. Inorg. Chem. 2014,
53, 12962.
(11) Keith, J. A.; Grice, K. A.; Kubiak, C. P.; Carter, E. A. J. Am.
Chem. Soc. 2013, 135, 15823.
(12) Bard, A.; Faulkner, L. Electrochemical Methods: Fundamentals and
Applications; 2nd ed.; John Wiley & Sons Inc.: New York, 2001.
(13) Smieja, J. M.; Sampson, M. D.; Grice, K. A.; Benson, E. E.;
Froehlich, J. D.; Kubiak, C. P. Inorg. Chem. 2013, 52, 2484.
(14) Smieja, J. M.; Kubiak, C. P. Inorg. Chem. 2010, 49, 9283.
(15) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456.
(16) Díez-Gonzal
874.
́
ez, S.; Nolan, S. P. Coord. Chem. Rev. 2007, 251,
(17) Tapu, D.; Owens, C.; VanDerveer, D.; Gwaltney, K.
Organometallics 2009, 28, 270.
(18) Valdes, H.; Poyatos, M.; Peris, E. Organometallics 2014, 33, 394.
́
(19) Voges, M. H.; Rømming, C.; Tilset, M. Organometallics 1999,
18, 529.
́
(20) Ullah, F.; Bajor, G.; Veszpremi, T.; Jones, P. G.; Heinicke, J. W.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.inorg-
chem.5b01698.
■
Angew. Chem., Int. Ed. 2007, 46, 2697.
S
(21) Gardner, B. M.; McMaster, J.; Liddle, S. T. Dalt. Trans. 2009,
6924.
(22) Bolotina, N. B. Crystallogr. Rep. 2007, 52, 647−658.
(23) Nielsen, D. J.; Cavell, K. J.; Skelton, B. W.; White, A. H. Inorg.
Chim. Acta 2006, 359, 1855.
(24) Andrew, R. E.; Chaplin, A. B. Dalton Trans. 2014, 43, 1413.
(25) Nielsen, D. J.; Cavell, K. J.; Skelton, B. W.; White, A. H. Inorg.
Chim. Acta 2002, 327, 116.
1H NMR spectra of proligand precursors, proligands, and
complexes, ATR-FTIR spectra of complexes, UV−vis
spectra of 1 and 2, cyclic voltammograms, DOSY NMR
data, crystallography details, plots of ip vs ν1/2 for
complexes, and DFT calculation sample input (PDF)
X-ray crystallographic data in CIF format (CIF)
(26) Danopoulos, A. A.; Tulloch, A. A. D.; Winston, S.; Eastham, G.;
Hursthouse, M. B. Dalt. Trans. 2003, 1009.
(27) Grundemann, S.; Albrecht, M.; Loch, J. A.; Faller, J. W.;
̈
Crabtree, R. H. Organometallics 2001, 20, 5485.
(28) Hahn, F. E.; Jahnke, M. C.; Gomez-Benitez, V.; Morales-
Morales, D.; Pape, T. Organometallics 2005, 24, 6458.
(29) Allen, F. H. Acta Crystallogr., Sect. B: Struct. Sci. 2002, 58, 380.
(30) Gambarotta, S.; Arena, F.; Floriani, C.; Zanazzi, P. F. J. Am.
Chem. Soc. 1982, 104, 5082.
AUTHOR INFORMATION
Corresponding Author
*E-mail: mwolf@chem.ubc.ca. Phone: (604) 822-1702. Fax:
(604) 822-2847.
■
(31) Schmidt, M. H.; Miskelly, G. M.; Lewis, N. S. J. Am. Chem. Soc.
1990, 112, 3420.
Notes
The authors declare no competing financial interest.
(32) Bhugun, I.; Lexa, D.; Savea
1769.
́
nt, J.-M. J. Am. Chem. Soc. 1996, 118,
ACKNOWLEDGMENTS
■
(33) Briegleb, G. Angew. Chem., Int. Ed. Engl. 1964, 3, 617.
(34) Goodnow, T. T.; Kaifer, A. E. J. Phys. Chem. 1990, 94, 7682.
(35) Aranzaes, J. R.; Daniel, M.; Astruc, D. Can. J. Chem. 2006, 84,
This research was supported by the Natural Sciences and
Engineering Research Council of Canada (NSERC). WestGrid
(www.westgrid.ca) and Compute Canada−Calcul Canada
(www.computecanada.ca) are acknowledged for access to
computational resources for DFT calculations. Prof. David
Wilkinson is thanked for use of the gas chromatograph. J.A.T.
thanks the NSERC for scholarship funding.
288.
(36) Rosen, E. L.; Varnado, C. D.; Tennyson, A. G.; Khramov, D. M.;
Kamplain, J. W.; Sung, D. H.; Cresswell, P. T.; Lynch, V. M.; Bielawski,
C. W. Organometallics 2009, 28, 6695.
(37) Kutt, A.; Rodima, T.; Saame, J.; Raamat, E.; Maemets, V.;
̈
̈
Kaljurand, I.; Koppel, I. A.; Garlyauskayte, R. Y.; Yagupolskii, Y. L.;
Yagupolskii, L. M.; Bernhardt, E.; Willner, H.; Leito, I. J. Org. Chem.
2011, 76, 391.
(38) Feigl, F.; Anger, V. Spot Tests in Organic Analysis, 7th ed.;
Elsevier Science: Amsterdam, The Netherlands, 1966.
REFERENCES
■
(1) Inglis, J. L.; MacLean, B. J.; Pryce, M. T.; Vos, J. G. Coord. Chem.
Rev. 2012, 256, 2571.
(2) Costentin, C.; Robert, M.; Savea
́
nt, J.-M. Chem. Soc. Rev. 2013,
(39) Arana, C.; Keshavarz, M.; Potts, K. T.; Abruna, H. D. Inorg.
̃
42, 2423.
Chim. Acta 1994, 225, 285.
(3) Appel, A. M.; Bercaw, J. E.; Bocarsly, A. B.; Dobbek, H.; DuBois,
D. L.; Dupuis, M.; Ferry, J. G.; Fujita, E.; Hille, R.; Kenis, P. J. A.;
Kerfeld, C. A.; Morris, R. H.; Peden, C. H. F.; Portis, A. R.; Ragsdale,
S. W.; Rauchfuss, T. B.; Reek, J. N. H.; Seefeldt, L. C.; Thauer, R. K.;
Waldrop, G. L. Chem. Rev. 2013, 113, 6621.
(40) Burfield, D. R.; Smithers, R. H. J. Org. Chem. 1978, 43, 3966.
(41) Williams, D. B. G.; Lawton, M. J. Org. Chem. 2010, 75, 8351.
(42) Noviandri, I.; Brown, K. N.; Fleming, D. S.; Gulyas, P. T.; Lay,
P. a; Masters, a F.; Phillips, L. J. Phys. Chem. B 1999, 103, 6713.
(43) Chai, J.-D.; Head-Gordon, M. Phys. Chem. Chem. Phys. 2008, 10,
6615.
(4) Mikkelsen, M.; Jørgensen, M.; Krebs, F. C. Energy Environ. Sci.
2010, 3, 43.
(44) Martin, J. M. L.; Sundermann, A. J. Chem. Phys. 2001, 114, 3408.
(5) Slater, S.; Wagenknecht, J. H. J. Am. Chem. Soc. 1984, 106, 5367.
(6) Hawecker, J.; Lehn, J.-M.; Ziessel, R. J. Chem. Soc., Chem.
Commun. 1984, 984, 328.
(45) Andrae, D.; Haußermann, U.; Dolg, M.; Stoll, H.; Preuß, H.
̈
Theor. Chim. Acta 1990, 77, 123.
K
DOI: 10.1021/acs.inorgchem.5b01698
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(46) Cossi, M.; Rega, N.; Scalmani, G.; Barone, V. J. Comput. Chem.
2003, 24, 669.
(47) Scheytza, H.; Rademacher, O.; Reißig, H.-U. Eur. J. Org. Chem.
1999, 1999, 2373.
(48) Hu, J.; Zhang, D.; Harris, F. W. J. Org. Chem. 2005, 70, 707.
L
DOI: 10.1021/acs.inorgchem.5b01698
Inorg. Chem. XXXX, XXX, XXX−XXX