Structural and electronic characterization of multi-electron reduced naphthalene (BIAN) cobaloximes

Article abstract of DOI:10.1039/c5dt00924c

Reported here are the syntheses and characterization of cobaloximes that feature a bis(imino)acenaphthene (BIAN) appended ligand. The X-ray crystal structures and spectroscopy (¹H NMR or EPR) of the complexes within the series [Co(aqdBF₂

Full text of DOI:10.1039/c5dt00924c

Dalton
Transactions
View Article Online
View Journal
PAPER
Structural and electronic characterization of
multi-electron reduced naphthalene (BIAN)
Cite this: DOI: 10.1039/c5dt00924c
cobaloximes†
Owen M. Williams, Alan H. Cowley and Michael J. Rose*
Reported here are the syntheses and characterization of cobaloximes that feature a bis(imino)ace-
naphthene (BIAN) appended ligand. The X-ray crystal structures and spectroscopy (1H NMR or EPR) of the
complexes within the series [Co(aqdBF2)2(MeCN)2] (1), [Co(aqdBF2)(MeCN)2]− (2) and [Co(aqdBF2)2-
(MeCN)2]2− (3, 3’) are reported and the 3-electron reduced complex [Co(aqdBF2)2(MeCN)2]3− (4) has been
prepared in situ and characterized by 1H NMR spectroscopy. The X-ray crystal structures revealed the
presence of a 6-coordinate CoII species (1), a 5-coordinate CoI species (2), and a 4-coordinate complex
(3, 3’). In the case of complex 3, evidence from single crystal EPR spectroscopy (g = 2.017, g⊥ = 1.987;
k
<10 G linewidths) in conjunction with DFT calculations indicate that the EPR signal originates from a de-
localized ligand-based unpaired spin. The frontier orbitals obtained from DFT calculations on 1, 2, 3, & 4
support the electronic assignments that were observed spectroscopically. The cathodic cyclic voltammo-
gram (CV) of the solvato congener in DMF solution, namely [Co(aqdBF2)2(DMF)2], exhibits three reversible
redox events near −1.0, −1.5 and −2.0 V vs. Fc/Fc+. Catalytic proton reduction was observed by CV near
the third redox peak. Compared with other cobaloximes (Ecat = −1.0 V), the delay of catalytic onset arises
from the existence of a series of resonance-stabilized intermediates.
Received 6th March 2015,
Accepted 1st April 2015
DOI: 10.1039/c5dt00924c
www.rsc.org/dalton
ated into heterogeneous systems such as glassy carbon,20
nickel oxide,21 carbon nanotubes,22 and gallium phos-
1. Introduction
The term molecular electrocatalysts broadly defines a category phide23,24 substrates. Furthermore, cobaloximes have
of homogenous catalysts, typically organometallic, that usually been explored in conjunction with homogeneous
require access to a reduced state prior to observable catalysis. photosensitizers,25–28 as components in aqueous systems,29–31
This type of catalysis is particularly well suited for kinetically as a functional catalyst on vesicle membrane surfaces,32 and
difficult transformations for which either reduction or oxi- used as part of an artificial hydrogenase in a hybrid SwMb
dation represents a limiting factor. The hydrogen evolution system.33 Moreover, increasingly elaborate ligand designs have
reaction (HER), oxygen evolution reaction (OER), and carbon been devised and studied. However, only marginal improve-
dioxide reduction are the most actively studied of these ments over the most basic complexes have been realized.34–36
reactions.1–6 Overall, cobaloximes remain one of the most well
studied and widely used classes of HER electrocatalysts.7,8
The two most widely known and cited cobaloxime case
studies are those that involve the dimethyl and diphenyl glyox-
Cobaloximes consist of a central cobalt atom that is ligated ime ligands (specifically, [Co(dmgBF2)2(MeCN)2] and
by the nitrogen donors of oxime ligands that bridge via hydro- [Co(dpgBF2)2(MeCN)2], Fig. 1). Synthesized originally by
gen bonds or chemically modified BF2 or BR2 bridges.9 The Schrauzer and Espenson,37,38 these complexes attracted atten-
various stages of cobaloxime catalysis have been investigated
by electrochemical methods,10–13
density functional
theory,14–16 electron paramagnetic spectroscopy,17 and photo-
physical techniques.18,19 Cobaloximes have also been incorpor-
Department of Chemistry, The University of Texas at Austin, Austin, TX 78712, USA.
E-mail: mrose@cm.utexas.edu
†Electronic supplementary information (ESI) available. CCDC 1052415–1052419.
For ESI and crystallographic data in CIF or other electronic format see DOI:
10.1039/c5dt00924c
Fig. 1 Cobaloximes studied in the literature and investigated herein.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
tion as electrocatalysts following seminal work by Peters,10,12
and Fontecave and Artero.11,13 In these early reports, the
relationship between the backbone substituents (i.e. phenyl vs.
methyl groups) showed that it is the potential of the observed
CoII/I redox couple that determines the potential at which the
HER reaction takes place. For example, in the case of the
fluorinated iron diglyoxime [Fe(dArFgBF2)2(py)2], backbone
tuning permitted a remarkable change in the FeII/I reduction
potential, while retaining an analogous catalytic cycle.39
Furthermore, a second redox peak that was detected in the
case of [Co(dmgBF2)2(MeCN)2] was deemed to be a ligand
based redox event.13 The redox catalysis in this case could be
achieved at either redox couple by an appropriate choice of
acid. However, the structure and electronics of this doubly
reduced species remain to be explored. As a consequence,
these reports stimulated our interest in incorporating a more
redox active backbone.
Bis(imino)acenaphthene (BIAN) ligands have proven to be
useful for a variety of catalytic systems, that range from C–C
bond coupling to polymerization.40 In particular, the
naphthalene backbone of this ligand facilitates a rich redox
chemistry.41–45 Moreover, the bis(diisopropyl)phenylimino
ligand (dpp-BIAN) is capable of undergoing four successive
one-electron reductions.41 Furthermore, the ease of prepa-
ration of BIAN complexes allows for the incorporation of a
wide variety of Schiff bases.46 Accordingly, we became inter-
ested in incorporating the BIAN motif into a known electro-
catalytic system. Overall, cobaloximes exhibit three attributes
Fig. 2 Cyclic voltammogram of 1·(DMF)2 obtained under an N2 atmos-
phere in a DMF electrolyte solution (0.1 M NBu4PF6). Experimental con-
ditions: 100 mV s−1; WE: GC, RE: Ag, CE: Pt. Inset: scan rate dependence
for each cathodic feature, indicating homogeneous processes in each
case.
consequence, the reduced species are formulated as [Co-
(aqdBF2)2(DMF)]−, [Co(aqdBF2)2]2−, and [Co(aqdBF2)2]3−. The
reducing equivalents in the latter two complexes are absorbed
by the ligand framework.
2.2. Chemical reductions
that render them desirable for a study of a BIAN appendage, To help probe this rich redox chemistry, a series of chemical
namely: (i) the capability for electrocatalytic activity, (ii) suit- reductions were performed with the objective of accessing the
able axial binding sites for the substrates, and (iii) a ligand putative complexes that had been detected in the cathodic
framework that is compatible with BIAN derivatization.
cyclic voltammogram. The stable, one-electron reduced species
The present work is focused on an analogous BIAN-cobalox- was produced via the reaction of the parent cobaloxime with a
ime, along with the X-ray crystal structures of several multi- near stoichiometric quantity of sodium naphthalide
electron reduced species. The foregoing examples were charac- (1.2 equivalents, E ≈ −3.10 V vs. Fc/Fc+)47 in THF, thus yielding
terized spectroscopically and also by means of DFT calcu- a stable bluish-violet solution. Re-dissolution of the solid
lations. An overview of the electrocatalytic studies (cyclic product in an MeCN solution containing excess 12-crown-4
voltammetry and bulk electrolysis) of this complex is also pre- afforded a violet colored solution. The crystalline product was
sented below.
isolated by vapor diffusion of Et2O into an MeCN solution to
afford the five-coordinate, square pyramidal product [Na(12-
crown-4)2][Co(aqdBF2)2(MeCN)] (2). Further details are pro-
vided in the X-ray crystallographic section. The difficulties that
were encountered in obtaining pure 2 in significant yields
prompted a search for an alternative method of generating the
2. Results
2.1. Electrochemistry
The cyclic voltammogram for the corresponding solvato singly reduced species. Regardless of stoichiometry, cobalto-
complex in DMF solution, namely 1·DMF2, is displayed in cene reduction (E ≈ −1.3 V vs. Fc/Fc+)47 resulted in the gene-
Fig. 2. This complex exhibits a reversible feature at −0.95 V vs. ration of a purple solution in acetonitrile, which was indicative
Fc/Fc+ (ΔEp ≈ 100 mV), thereby generating [Co(aqdBF2)2- of a one electron reduction. The assignment was corroborated
(DMF)]−. This response is almost identical to that of the CoII/I on the basis of single crystal XRD (Fig. S2†) and 1H NMR
reduction potential that was reported for [Co(dmgBF2)2(DMF)2] spectroscopy.
(E1/2 ≈ −1.0 V vs. Fc/Fc+).13 The ensuing cathodic sweep
The two-electron reduced species was isolated by treatment
revealed two additional reversible features at −1.56 and of a solution of the parent cobaloxime 1 with an excess potass-
−2.00 V vs. Fc/Fc+ (ΔEp ≈ 60 mV in both cases). Each feature ium graphite (>3 equivalents, E = −2.90 V vs. Fc/Fc+)47 in THF,
exhibits a linear increase in current with respect to scan rate thus forming a dark navy blue solution that was markedly
(Fig. 2, inset; see Fig. S1† for full CV traces), thus confirming different from the violet colored solution that had been
the occurrence of homogeneous processes in each case. As a observed for the corresponding one-electron reduced complex.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
Extraction of the product into a MeCN solution containing 2.3. X-ray crystal structures
dibenzo-18-crown-6
afforded
[K(dibenzo-18-crown-6)]2-
[Co(aqdBF2)2] (3), which features the presence of a cobalt
2.3.1 Ligand (aqdH2). The X-ray crystal structure of the
center in a square planar geometry (vide infra, X-ray Crystallo- parent ligand, acenaphthenequinone dioxime (aqdH2,
graphy section). A similar procedure was found to be effective Fig. S3†) exhibited structural parameters that are consistent
for the isolation of the alkyl crown congener 3 of formula with those of other BIAN-Schiff base ligands. Moreover, the
[K(18-crown-6)]2[Co(aqdBF2)2]. Finally, the preparation of the CvN bond lengths of 1.285(2) and 1.286(2) Å are consistent
three-electron reduced compound (4) was carried out by the with the presence of double bond character and the C–C bond
treatment of 1 with an excess of Na/Hg (E ≈ −2.4 V vs. Fc/Fc+) length of 1.502(3) Å is in agreement with the value expected
in a THF solution in the presence of an excess of 12-crown-4. for a single bond. The N1–C1–C2–N2 torsion angle of 6.6(2)° is
It is important to note that the green solution is not most likely due to the constrained five membered ring of the
observed in the absence of a supporting crown ether. The acenaphthene framework. All other bond distances fall within
role of the alkali-crown complex in stabilizing the the expected ranges.
reduction state of the complex is highlighted in the X-ray
crystallography section.
2.3.2 [Co(aqdBF2)2(MeCN)2] (1). The X-ray structure of 1
(Fig. 4) revealed the presence of a CoII ion in a pseudo-octa-
Overall, an evaluation of the reduction potentials and stoi- hedral geometry. This outcome was anticipated for a cobalox-
chiometries of the reductants that were employed failed to ime of this type that had been crystallized from a coordinating
reveal a coherent trend regarding the isolated products solvent (MeCN). Furthermore, the axial MeCN ligands are
(among NaC10H8, KC8, Na/Hg, all have a reduction potential tilted away from a linear orientation by 14.8°, and each MeCN
cathodic of −2.0 V vs. Fc/Fc+). The CoCp2 reduction clearly molecule is flanked by a BF2 bridge that is tilted up or down
shows a dependence on potential since it only leads to the out of the plane of the macrocycle in order to avoid the axial
singly reduced product regardless of stoichiometry. Inspection ligand. The structure of the aqdBF2 ligand of 1 is essentially
of the X-ray crystal structures (vide infra) revealed the presence planar, as evidenced by the torsion angle of 1.4(1)° for the C2–
of secondary effects due to the interaction of the alkali metals C1–N1–Co1 chain. The Co–Nox bond distances of 1.896(1) Å for
with the complex that represent a determining factor in terms 1 (fourfold crystallographic symmetry) are slightly longer than
of the identity of the final product. For example, reduction of those found for closely related cobaloximes [Co(dmgBF2)2-
the parent complex with NaC10H8, KC8, or Na/Hg in neat THF (MeCN)2] (1.878(1), 1.881(2) Å)17 and [Co(dpgBF2)2(MeCN)2]
each generate blue solutions that are indicative of the for- (1.881(3), 1.889(3) Å).12 The foregoing results are likely to be a
mation of the 2-electron reduced species. However, subsequent consequence of the expanded NvC–CvN bite angle in the
extractions of the various alkali salts into MeCN–crown ether fused naphthalene ring system, which results in N–Co–N
mixtures (to aid solubility and crystallization) produced angles of 84.66(4)° for aqd, versus 81.70(7)° for dmg or 81.8(1)°
different results. For example, dissolution of the Na salt of the for dpg. The axial Co–NMeCN distances of 2.2712(2) Å are also
2-electron reduced blue species in an MeCN–12-crown- longer than those found in the dmgBF2 and dpgBF2 congeners
4 mixture caused an immediate color change from blue to (2.253(2), 2.241(3) Å). However, the CvN and C–C bonds in
violet, (which is indicative of an auto-oxidation process) and the macrocycle (1.293(1) and 1.488(2) Å) are commensurate
resulted in the formation of the 1-electron reduced species 2. with those of the neutral ligand (1.286(2), 1.285(2) and
By contrast, extraction of the K salt of the 2-electron reduced 1.502(2) Å). Furthermore, the acene C–C bond distance in 1
blue species into a MeCN solution that contained 18-crown-6 (1.488(2) Å) is very close to that observed for the free ligand
derivative retained the blue color, which subsequently pro- (1.502(2) Å), thus indicating that no electronic change or sig-
ceeded to generate the authentic crystalline samples of the nificant delocalization of charge from the cobalt(II) center to
2-electron reduced species 3.
the ligand framework is observable by XRD.
Interestingly, a distinct reductive behavior was evident in
2.3.3 [Na(12-crown-4)2][Co(aqdBF2)2(MeCN)] (2). The cobalt
THF solution that depended on the presence or absence of a center in this species adopts a square based-pyramidal struc-
crown ether. In contrast to the blue solutions that were appar- ture, which effects a 0.415 Å displacement of the cobalt ion
ent in the absence of the crown ether, the reduction of 1 with out of the equatorial plane and a significant contraction of the
Na/Hg in a THF solution that contained 12-crown-4 resulted in single axial Co–NMeCN bond to 2.007(3) Å, as compared with
the formation of a green solution, which was indicative of the the two axial Co–NMeCN bonds (2.2714(16) Å)) that were
presence of the 3-electron reduced species 4. Overall, the X-ray observed in the case of the parent complex. A comparison of
crystal structure data revealed the presence or absence of the structural parameters of 2 with those of the reduced diphe-
site-specific interactions between the crowned Na+ or K+ ions nyl (dpg) cobaloxime analog [Na(12-crown-4)2][Co(aqdBF2)2-
with the oxime-N/O or BF2 moieties. Evidently, it is this (MeCN)] reveals that the bond lengths are very similar (N–Coavg
counterion cooperativity that represents a critical factor in 1.851 Å; CvNavg 1.321 Å; C–Cavg 1.457 Å).12 However, within
terms of stabilizing the reduced states during isolation. the ligand framework there is a slight contraction of the N–Co
Furthermore, it is clear that careful attention must be paid bond length (∼0.016 Å), an elongation of the N–C bond length
to the presence (or absence) and identity of the supporting (∼0.017 Å), and a contraction of the C–C bond length
alkali-crown species.
(∼0.040 Å) in comparison with that of 1. The slight differences
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 3 Synthetic route for the ligand aqdH2, 1 and the chemically reduced species 2, 3 and 4.
2.3.4 [K(dibenzo-18-crown-6)]2[Co(aqdBF2)2] (3). The struc-
ture of the doubly reduced complex, 3 (Fig. 5, center) revealed
an increase of the planarity of the entire macrocycle in
comparison with that of the singly reduced (Na salt) structure
(24.01 cf. 47.52 degrees for the mean plane angle between the
naphthalene backbone units). This outcome is likely a conse-
quence of the orbital overlap required for effective delocalization
of the additional electron. The cobalt center in this compound
adopts a square planar structure, and is accompanied by a
further contraction of the N–Co bond length (∼0.024 Å), an
elongation of the N–C bond length (∼0.030 Å), and a lengthen-
ing of the C–C bond (∼0.012 Å) in comparison with those of
the singly reduced species. Of particular note are the short
contacts between the two chelated potassium cations and the
fluorine atoms (K⋯F = 2.621(2), 2.655(2) Å) on the BF2 moi-
eties (Fig. 5, bottom center). The open-face chelation of the K+
ion by dibenzo-18-crown-6 provides coordination sites for the
direct interaction of the alkali metal with the electronegative
moieties on the cobalt macrocycle. In contrast, the bis-chelated
nature of the sodium counterion [Na(12-crown-4)2] found in 2
precluded any direct interaction between the alkali metal and
Fig. 4 ORTEP diagram (50% thermal ellipsoid plots) of the parent
complex [Co(aqdBF2)2(MeCN)2] (1). All hydrogen atoms have been
omitted for clarity. Selected bond lengths (Å) and angles (°): N1–Co1,
1.8961(9); N2–Co1, 2.271(2); C1–N1, 1.293(1); C1–C1*, 1.488(2); N1–
Co1–N1*, 84.66(4).
in the C–N and C–C bond distances between 1 and 2 may indi- the cobaloxime. As noted in the Chemical Reductions section
cate partial delocalization of the reducing electron onto the (vide supra), it can be inferred that such a direct contact with
N–C–C–N fragment. The bond lengths, however, show altered the K+ ion in 3 (and 3′, see below) stabilizes the reduction state
parameters when compared with that of a model one-electron of the complex thereby preventing auto-oxidation (see Fig. 3).
reduced BIAN, namely the sodium complex of diiso-
2.3.5 [K(18-crown-6)]2[Co(aqdBF2)2] (3′). In
a
similar
propylphenyl-BIAN (dpp-BIAN). In comparison with Na-dpp- fashion to 3, the 18-crown-6 ligated K+ salt, 3′ displayed the
BIAN, complex 2 exhibits slightly shorter CvN (∼1.310 same square planar geometry at the cobalt center, with close
cf. 1.328 Å) and slightly longer C–C (∼1.448 cf. 1.446 Å) bond contact between the crown-stabilized cation and the cobalox-
distances.41 However, due to the significant change in the ime (Fig. 5, bottom right). Examination of the structure of 3′
coordination environments of 1, 2, and [Na][dpp-BIAN] it is revealed that the potassium cations interact with the macro-
difficult to assign formal ligand charges and/or metal oxi- cycle in a different manner, such that each [K(18-crown-6)]+
dation states based solely on structural data.
unit adopts close contacts with a lateral set of two N–O frag-
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
Fig. 5 X-ray crystal structures of 2, 3, & 3’ (50% probability thermal ellipsoids). (top) Only the cobalt-containing anions are displayed, and hydrogen
atoms, counterions, and solvent molecules have been omitted for clarity. (bottom) Full structures, cation inclusive are shown below. Selected bond
lengths (Å) and angles (°), with symmetry generated atoms are designated by an asterix (*): [CoI(aqdBF2)2(MeCN)]−: N1–Co1, 1.879(2); N2–Co1,
]
1.887(2); C1–N1, 1.311(2); C8–N2, 1.308(2); C1–C1*, 1.439(3); C8–C8*, 1.445(3). [CoI(aqdBF2)2•− 2−, dibenzo-18-crown-6 potassium salt: N1–Co1,
1.863(2); N2–Co1, 1.862(2); N3–Co1, 1.859(2); N4–Co1, 1.854(2); C1–N1, 1.339(3); C2–N2, 1.333(4); C13–N3, 1.344(3); C14–N4, 1.335(4); C1–C2,
1.460(4); C13–C14, 1.455(4). [CoI(aqdBF2)2
C1–C2, 1.326(3).
]
•− 2−, 18-crown-6 potassium salt: N1–Co1, 1.847(2); N2–Co1, 1.834(2); C1–N1, 1.328(3); C2–N2, 1.326(3);
ments (an η4 arrangement) on the macrocycle (K⋯O = 2.896(2), ion in 1 supported by the planar (aqdBF2)2 macrocycle, and co-
3.121(2) Å; K⋯N = 3.160(2), 3.400(2) Å). The overall closer ordinated by the two axial MeCN ligands. The simulated EPR
approach of the entire [K(18-crown-6)]+ fragment to the cobalox- signal aligned well with that of the experimental spectrum,
ime that was observed in 3′ is a consequence of the less bulky, thus corroborating the features that were apparent in the range
alkyl framework of the crown in 3′ compared with that of the of g ≈ 2–2.3 (gx = 2.29, gy = 2.20, gz = 2.02). The result is also
dibenzo-crown framework in 3. Furthermore, the planarity of 3′ consistent with the solid state magnetic susceptibility
was found to have increased in comparison with that of its con- measurement that was performed on a crystalline sample of
gener, 3 (N–Co–N ∼180° cf. 173°). Interestingly, the increased 1 (μeff = 1.71μB, 298 K; predicted μSO = 1.73μB). Overall, the
planarity of 3′ occurs in conjunction with an ‘anti’ orientation EPR spectrum of 1 is very similar to those reported for
of the BF2 groups, whereas the BF2 linkers in 3 adopt a ‘syn’ [Co(dmgBF2)2] in similar environments (gx = 2.26, gy = 2.17,
conformation. The relevant metrical parameters for this gz = 2.01 Az = 338 MHz).17 However, the anticipated 59Co (I =
complex are comparable to those that were found for 3, albeit 72) hyperfine pattern was not evident in the solid state
with slightly decreased N–Co bond lengths (∼0.018 Å), shorter measurement. This outcome is most likely due to the close
CvN bonds (∼0.016 Å), and acenaphthene C–C bonds proximity of the paramagnetic centers in the crystalline state.
(∼0.015 Å). The discrepancies are attributable to the planar However, in a frozen solution of MeCN–THF (10% v/v) at 85 K,
orientation of the ligating moieties.
complex 1 exhibited a significantly more resolved spectrum
(Fig. 6, bottom panel) with primary features at gx = 2.31, gy =
2.26, gz = 2.02. The latter outcome is also indicative of the pres-
ence of a low-spin d7 ion in solution; no features near g ≈ 4.3
were observed. Furthermore, the solution phase EPR spectrum
of 1 exhibited the anticipated eight-line pattern for the I = 72
59Co nucleus, with a coupling constant of Az = 110 G
(∼310 MHz). The correlation of the hyperfine splitting pattern
2.4. Spectroscopic characterization
2.4.1 EPR of [Co(aqdBF2)2(MeCN)2] (1). The EPR spectrum
of crystalline 1 was obtained by means of a solid state powder
measurement at 85 K, the overall result of which is displayed
in the upper panel of Fig. 6. This spectrum is indicative of the
presence of a low-spin, axial system as expected for the d7 CoII
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
Fig. 7 1H NMR spectrum of [CoCp2][Co(aqdBF2)2(MeCN)] (2’) in CD3CN
at 298 K (400 MHz). The complex was prepared from cobaltocene and
structurally characterized as the cobaltocenium salt (Fig. S2†).
gous reduction was performed with cobaltocene, which
afforded the cobaltocenium salt, 2′. The prevalent signals
emanate from the one electron reduced species, and exhibit
resonances that correspond to those of the cobaloxime anion
and the cobaltocenium cation. Identification of the anion as
the desired product was confirmed by means of a single
crystal X-ray diffraction study (Fig. S2†). The anion that
resulted from this one-electron reduction afforded a stable
1
CD3CN solution of the desired complex (Fig. 7). The H NMR
spectrum of this cobaltocenium salt revealed the presence of a
diamagnetic species that exhibited two doublets (δ = 8.15 ppm,
J = 6.8 Hz; δ = 7.92 ppm; J = 8.0 Hz) along with a doublet of
doublets (δ = 7.56 ppm, J = 6.8, 8.0 Hz) in the aromatic region,
thus confirming the presence of the naphthalene backbone. In
comparison with the 1H NMR spectrum of the free aqdH2
ligand (δ = 12.36 ppm, singlet; δ = 8.35 ppm, doublet, J =
6.6 Hz;δ = 8.01 ppm, doublet, J = 7.8 Hz; δ = 7.70 ppm, doublet
of doublets, J = 6.8, 8.0 Hz), these resonances show only a
slight shift in the aromatic region, signifying that little to no
change had taken place in the electronic environment of the
naphthalene rings.
Fig. 6 X-band EPR studies of [Co(aqdBF2)2] (1). Top Panel: (Black line)
powder EPR spectrum of a solid sample of 1 at 85 K; (red line) simulated
EPR spectrum; simulation parameters: gx = 2.29, gy = 2.20, gz = 2.02.
Bottom panel: X-band EPR spectrum of 1 in a frozen glass of 10%
MeCN–THF (v/v). Instrument parameters, powder spectrum: frequency,
9.40 GHz; modulation, 100 kHz; power, 0.2 mW; field modulation,
2 G. Instrument parameters, solution spectrum: frequency, 9.44 GHz;
modulation, 100 kHz; power, 25 mW; field modulation, 10 G.
2.4.3 EPR of [K(dibenzo-18-crown-6)]2[Co(aqdBF2)2] (3).
Solution phase EPR measurements of 3 in either frozen MeCN
or MeCN–tol glass (85 K) revealed a broad feature (∼100 G line-
width) centered near g ≈ 2.0 (Fig. 8, top). In contrast, the
single crystal EPR spectrum of 1 at 85 K (Fig. 6, bottom) fea-
tured a single, extremely sharp peak (linewidth <10 G) with a
g-value that systematically ranged between 1.987 and 2.017
depending on the orientation of the crystal. The intensity of
this feature increased as the resonance shifted to lower mag-
netic fields upon rotation of the sample tube. Furthermore, the
intensity increased approximately two-fold, thus supporting the
with the gz = 2.02 feature is consistent with the primary orien-
tation of the unpaired spin being aligned along the z axis, as
expected for the singly occupied dz2 orbital. This interpretation
is also consistent with the long Co–NMeCN bond distances
(2.2714(16) Å) that were evident in the crystal structure of 1.
2.4.2 1H NMR of [CoCp2][Co(aqdBF2)2(MeCN)]− (2′).
Following the isolation as the Na-crown salt, the re-dissolution
of 2 in CD3CN resulted in slow bleaching of the solution. This
outcome is possibly due to interaction of the residual moisture
from the CD3CN solvent with the reduced species, which, in
turn could result in hydrolysis of the complex. The decompo-
sition of the complex in CD3CN is not due to an inherent
instability or reactivity of the complex since 2 was easily
extracted into and crystallized from CH3CN during the iso-
lation process. In order to circumvent this limitation, an analo-
assignments of g⊥ = 2.017 and g = 1.987 and indicating the
k
presence of an axially anisotropic radical. The proximity of this
single feature to the free electron value (g = 2.023), along with
the extremely narrow linewidth, strongly suggested the presence
of a primarily organic radical. The discrepancy between the line
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
in the aromatic region. As expected for the acenaphthene
framework, a pattern of two doublets (δ = 8.21 ppm, J = 7.6 Hz;
δ = 7.61 ppm, J = 8.4 Hz) and one triplet (δ = 7.44 ppm, J =
8.0 Hz) was observed. Overall, these resonances are quite
similar to those of the one-electron reduced species 2′, which
exhibits such features at δ = 8.15, 7.92 and 7.56 ppm.
2.5. DFT Calculations
Density functional theory calculations were performed with
the view of confirming the electronic structures that had been
proposed for the reduced cobaloxime complexes (Fig. 9). The
geometry optimizations were performed using the X-ray struc-
ture coordinates of 1, 2, and 3. The optimized structure of 1 (S =
1/2) closely mimics that of the crystal structure, albeit with a
slight reorientation of the axial acetonitrile solvent. The fron-
tier orbital (SOMO) of this particular compound showed evi-
dence of the presence of an unpaired spin located in the dz2
orbital of the metal center. Accordingly, this result confirmed
the assignment that had been made based on the EPR results
that were discussed earlier.
The frontier orbital (HOMO) of 2 shows a pairing of the
electrons in the dz2 orbital. This in turn accounted for the dia-
magnetism observed for the one electron reduced complex.
There is clear evidence for overlap between this orbital and the pz
orbitals located on the nitrogen atoms. This overlap may account
for the changes in bond distances that were detected in the X-ray
crystal structure. Furthermore, the location of the additional elec-
tron confirms that only a small change had taken place in the
electronic environment of the acenaphthene backbone.
The frontier orbital (SOMO) and spin density plots
(Fig. S5†) that were obtained for 3 indicate that the unpaired
spin is located in an orbital that spans the metal center and is
primarily delocalized across the two N–C–C–N fragments (S =
12). Interestingly, the spin density also extended slightly onto
the glyoximato-O. An extended view of the spin density
revealed that the spin density encapsulates the central cobalt
ion, but is only slightly localized on the metal center. By con-
trast, at this isosurface value the ligand framework is replete
with spin density. Unrestricted DFT calculations showed a
similar alpha spin density plot, albeit accompanied by signifi-
Fig. 8 EPR studies of [K(dibenzo-18-crown-6)]2[Co(aqdBF2)2] (3): (top)
solution EPR spectrum obtained in frozen MeCN at 85 K; (bottom) single
crystal EPR of 3 obtained at 85 K during a 90° rotation using a controlled
angle goniometer. Instrument parameters, solution spectrum: fre-
quency, 9.44 GHz; modulation, 100 kHz; power, 20 mW; field modu-
lation, 0.1 G. Instrument parameters, single crystal spectra: frequency,
9.45 GHz; modulation, 100 kHz; power, 10 mW; field modulation, 0.1 G.
widths of the single crystal and frozen-solution measurements
can be accounted for by assuming random orientations (and
an average of the signals) in the case of the frozen solution.
This is further corroborated by a polycrystalline measurement,
which showed a complex spectrum, with no discernible
pattern upon rotation (see Fig. S4†). The assignment of the
additional electron as an organic based radical was supported
by DFT calculations that were carried out on complex 3
(vide infra, DFT Calculations section).
2.4.4 1H NMR spectrum of [Co(aqdBF2)2]3− (4). The three-
electron reduced compound 4 was characterized on the basis
of 1H NMR spectroscopy. This complex was generated in situ in
d8-THF solution by reduction with Na/Hg in the presence of
12-crown-4. This reaction resulted in the formation of a green
1
solution that could be analyzed (298 K) by H NMR. The dia-
magnetism of 4 was confirmed by the presence of sharp peaks
Fig. 9 DFT calculated frontier orbitals for 1–3 (left to right) using the B3P86 functional in conjunction with a Wachters/TZVP hybrid basis set; geo-
metry optimizations were performed with a smaller basis set (see Experimental). The surfaces are displayed at the 0.05 isosurface value. Spin density
plots for 3 are available in the ESI.†
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
2
2
cant spin contamination (Scalc ≈ 1.00; Sexp = 0.75). This may
indicate that contributions from both CoIL•− and CoIIL2− are
feasible for this particular compound.
2.6. Electrocatalysis
The proclivity of the title complex for proton reduction was
also explored. For this purpose, CV titrations were performed
along with additions of two commonly used proton
sources, namely p-toluenesulfonic acid [TsOH is dissociated,
the functional acid is DMF-H+] and benzoic acid (PhCO2H).
The overpotentials for H+ reduction were determined by the
first derivative method described by Fontecave and Artero.48
Interestingly, the addition of TsOH failed to result in any
increase in current at the CoII/I redox couple, a process that is
otherwise observed for the majority of cobaloximes. Alterna-
tively, low concentrations of TsOH (2.5 mM) effect a significant
increase in current at the third redox event (imax/ip = 5.77),
which is one full volt cathodic of the CoII/I redox potential. At
further cathodic potentials (−2.5 V), a related hydritic cobalt
intermediate (tentatively assigned) also displayed a reversible
peak. At higher concentrations of TsOH (up to 25 mM), the cata-
lyzed 2H+ → H2 transformation was evidenced by a significant
increase in current (imax/ip = 43.3, 25 mM, Fig. 10) at −2.3 V vs.
Fc/Fc+, which corresponds to a rather high overpotential (ηTsOH
=
1.5 V, Fig. S6†). By contrast, when [Co(dmgBF2)2(DMF)2] was
monitored with a similarly strong acid (HBF4) in DMF a low over-
potential of η = 0.4 V was determined.12
An analogous set of experiments with the weaker acid
PhCO2H (10 mM) revealed a surprisingly similar increase in
current at the third redox event (imax/ip = 6.06). However, a cat-
Fig. 10 CV studies of 1·(DMF)2 in the presence of a weak acid (benzoic
acid, top) or in the presence of a strong acid (tosic acid, bottom). Insets
show the I–V response in the presence of a 25 mM concentration of the
respective acid using either a glassy carbon (black dashed) or platinum
(color dashed) button electrode. Standard experimental conditions:
100 mV s−1; WE: GC, RE: Ag, CE: Pt.
alytic wave was not observed until −2.5 V vs. Fc/Fc+ (imax/ip
=
20.0, Fig. 10). The acid concentration dependence for PhCO2H
is indicative of homoconjugation (ηPhCO H = 1.1 V, Fig. S7†) as
2
expected for the presence of a carboxylic acid in an organic
solvent. The linear dependence of the scan rate versus peak
current for the CVs that were carried out in the presence of
PhCO2H suggests that these processes are homogeneous
(Fig. S8†).
of some of the complexes that have been used in bulk electro-
lysis experiments have been called into question recently.50,51
Bulk electrolysis (BE) experiments were attempted in a
custom H-cell (Fig. S9†). In any given experiment, a large
amount of current was found to pass during the electrolysis,
and the hydrogen production was detected by GC. The sub-
3. Conclusions
1 The BIAN redox reservoir acts as an effective electron sink
sequent rinse tests, however show a similar efficiency for the during multiple-electron reduction.
used electrodes. This similarity of currents can be attributed to
2 During isolation of the reduced species with alkali-based
native electrode surface reactions, catalyst degradation, metal reductants, there is a critical dependence on the presence
deposition, and/or solvent degradation. Such processes are versus absence, as well as the identity of the crown ether
more prevalent over the extended time periods associated with employed. In general, complete sequestration of the alkali
bulk electrolysis. The extreme potentials required to observe metal from the reduced complex results in auto-oxidation pro-
(by CV) a catalytic response render the analysis of catalytic cesses during extraction and crystallization.
capability difficult. While no comprehensive study of acids has
3 In contrast, crown-ethers that allow for direct contact of
been performed in DMF as in the case of MeCN,49 it is likely the alkali metal ion with Lewis basic sites on the complex
that these experiments are being performed near or beyond facilitate isolation of the doubly reduced states.
the edge of the useful potential window (EBE < Einf.) for each
4 The capacity of the BIAN reservoir for electron storage allows
acid. In comparison with the literature cobaloximes, the limit- for a clean, sequential progression through paramagnetic → dia-
ations regarding the operating potential for proton reduction magnetic → paramagnetic → diamagnetic species (1 → 2 → 3 → 4)
in this particular case are evident. Furthermore, the stabilities as judged by EPR and NMR experiments.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
5 The 2-electron reduced species exhibits a primarily
4.2.2 [Na(12-crown-4)2][Co(aqdBF2)2(MeCN)] (2). Complex
organic radical that is delocalized into the BIAN glyoxime 2 was isolated by dissolving 20.7 mg of the parent compound
ligand. (0.0314 mmol) in 8 mL of THF. Following this, 1.29 equiva-
6 CV experiments demonstrate that the DMF solvato species lents of Na[C10H8] was added to the reaction mixture from a
of 1, namely 1·(DMF)2, catalyzes the reduction of H+ to H2 standard solution (0.3 mL, 0.135 M Na[C10H8] in THF). The
(confirmed by GC detection in BE). However, the negative latter solution was allowed to stir for thirty minutes ultimately
potentials associated with this catalytic activity and the associ- generating a purple-blue solution. The solvent was removed
ated H+ reduction activity of the glassy carbon rod WE pre- under reduced pressure and the resulting powder dissolved in
clude our ability to discern catalysis by the complex from the acetonitrile, affording a violet solution with a brown precipi-
electrode in bulk electrolysis measurements.
tate, which was removed by filtration. Vapor diffusion of Et2O
7 The available resonance forms between the ligand frame- into the latter solution in the presence of excess (100 mg,
work and the cobalt metal center prevent proton reduction at 0.568 mmol) 12-crown-4 yielded X-ray quality crystals of com-
potentials that are usually observed for cobaloximes.
pound 2. Yield: 10.5 mg; 0.011 μmol; 34%. IR (cm−1): 2868 (w),
8 Overall, the potential utility of the BIAN fragment to serve 1609 (w), 1519 (w), 1484 (w) 1443 (w), 1411 (w), 1363 (w),
as an electron reservoir motif for multi-electron catalysis is 1304 (w), 1244 (w), 1143 (m), 1093 (s), 983 (s), 912 (s), 849 (m),
evident.
827 (m), 789 (m). UV/Vis in THF–MeCN (10 : 1), λmax (nm): 560,
770. This complex was unstable upon re-dissolution in CD3CN.
1
An alternative salt (below) was prepared to obtain a H NMR
spectrum of the singly reduced anion.
4. Experimental
4.2.3 [Co(Cp)2][Co(aqdBF2)2(MeCN)] (2′). The parent coba-
loxime (23.7 mg, 0.036 mmol) was dispersed in MeCN solu-
tion. The subsequent addition of 21.3 mg of cobaltocene
(0.126 mol) resulted in the generation of a clear violet solution.
The latter solution was allowed to stir for thirty minutes, fol-
lowing which, most of the MeCN solvent was removed under
reduced pressure. The solution was then subjected to the
addition of Et2O, generating a purple precipitate of 2′. The
resulting filtrate was dissolved in CD3CN in order to obtain a
satisfactory 1H NMR spectrum (Fig. 7). Yield: 25 mg;
0.031 mmol; 86%. Crystallization of this compound was
effected by vapor diffusion of Et2O into the above-mentioned
solution to afford X-ray quality crystals. 1H NMR: δ = 8.15 ppm,
d; J = 6.8 Hz; δ = 7.92 ppm, d; J = 8.0 Hz, δ = 7.56 ppm, dd; J =
4.1 Reagents and procedures
Cobalt(II) acetate, potassium graphite, sodium metal, mercury,
and cobaltocene were purchased from Strem Chemicals. Ace-
naphthenequinone, hydroxylamine hydrochloride, boron tri-
fluoride etherate, naphthalene, dibenzo-18-crown-6, 18-crown-
6 and 12-crown-4 were purchased from Acros Organics and
used as received. The ligand aqdH2 was synthesized according
to the published procedure of Satake et al.52 All solvents used
in the syntheses were purchased from Fisher Scientific, and
dry solvents were obtained using a Pure Process Technology
solvent purification system. Anhydrous DMF was purchased
from Acros Organics and used without further purification for
all the electrochemical studies.
6.8, 8.0 Hz;
δ = 5.52 ppm, singlet. Elem. Analys. for
C36H25B2Co2F4N5O4 (%) calcd: C 53.57, H 3.12, N 8.68; found:
C 53.69, H 3.08 N 8.62.
4.2 Synthesis of metal complexes
4.2.1 [Co(aqdBF2)2(MeCN)0] (1). Complex 1 was prepared
4.2.4 [K(dibenzo-18-crown-6)]2[Co(aqdBF2)2] (3). Complex
by using a method that is similar to that described by 3 was synthesized by dissolving 22.4 mg of the parent com-
Bakac and Espenson.38 A 50 mL Schlenk flask was charged pound (0.034 mmol) in 8 mL of THF. After the addition of 3.2
with (0.50 g, 2.8 mmol) of acenaphthenequinone dioxime and equivalents of KC8 (15 mg; 0.111 mmol), the resulting solution
(0.30 g; 1.2 mmol) of Co(OAc)2·4H2O, and placed under was allowed to stir for thirty minutes and filtered to remove
vacuum for 20 min. Dry Et2O (30 mL) was added, thereby gen- any remaining traces of the reductant. To this solution
erating a light brown slurry. Next, 1.5 mL of BF3·Et2O was 3.7 equivalents of dibenzo-18-crown-6 (45 mg; 0.125 mmol)
added via syringe. After stirring the reaction mixture for were added, resulting in a slight color shift from deep
12 hours, the resulting dark brown precipitate was collected by navy blue to a dark blue-green. The solvent was then removed
subsequent filtration. The brown/red powder that had formed under reduced pressure and the resulting powder was dis-
was stirred in 20 mL of MeCN for one hour. Subsequent fil- solved in acetonitrile, which retained the blue-green hue
tration resulted in the isolation of a bright red powder of 1: observed in THF solution. Vapor diffusion of Et2O into the
0.249 g; 0.378 mmol; 31.5%. IR (cm−1): 3603 (m), 3530 (m), latter solution afforded a crop of X-ray quality crystals of the
1623 (m) 1484 (w) 1416 (w), 1134 (s), 987 (s), 965 (s), 911 (s), desired compound. Crystalline Yield: 21 mg; 0.015 mmol;
826 (s), 775 (s), 539 (m), 508 (m). UV/Vis in THF, λmax (nm): 45%. IR (cm−1): 2929 (w), 1587 (m), 1504 (s), 1452 (m) 1248 (s),
505, 340, 280 μeff = 1.71 μB. EPR: gx = 2.29, gy = 2.20, gz = 2.02, 1213 (s), 1124 (s), 993 (s), 942 (s), 813 (m), 766 (m), 743 (s). UV/
85 K, solid powder (Fig. 6). Elem. Analys. for Vis in THF, λmax (nm): 605 (broad), 770 (broad). EPR: see text
C28H18B2CoF4N6O4 (%) calcd: C 51.03, H 2.75, N 12.75; found: and Fig. 8. This complex proved too unstable for extended
C 50.98, H 2.63 N 12.37. HR-MS for [1–2MeCN + Na]+: calcd, times in the solid state to obtain satisfactory elemental
600.02130; found, 600.02090.
analysis.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
4.2.5 [K(18-crown-6)]2[Co(aqdBF2)2] (3′). The complex 3′ working electrode and counter electrode. The silver wire
was synthesized by a similar procedure as for 3, with 21.6 mg reference electrode was connected to the potentiostat by
of 1 (0.033 mmol) and 3.1 equivalents of KC8 (13.7 mg; means of a copper wire. In the H-cell experiments, a pre-
0.102 mmol). Following filtration to remove the reducing charge period of approximately three hours was employed by
equivalents, excess 18-crown-6 was added to the solution holding the potentiostat to a potential just cathodic of the
(34.9 mg; 0.132 mmol). Crystallization of this compound was second redox peak, thus generating an intense blue solution of
achieved directly from the above THF solution at reduced the pre-catalyst.
temperatures (−20° C) in an argon atmosphere. Crystalline
Yield: 16 mg; 0.014 mmol; 41%. IR (cm−1): 2882 (m), 1601 (m),
1520 (m), 1473 (w) 1410 (w), 1352 (m), 1250 (w), 1109 (s), 1060
4.4 X-ray crystallography
(m), 1008 (s), 825 (m), 778 (s). This complex proved too
unstable for extended times in the solid state to obtain satis-
factory elemental analysis.
Structures were solved, refined, and analyzed using standard
procedures53–57 Definitions used for calculating R(F), wR(F2)
and the goodness of fit, S, are given below.58 All figures were
generated using Mercury.59 Details regarding the data collec-
tion and solution parameters for aqdH2 and the cobalt com-
plexes are summarized in Table 1.
4.3 Electrochemistry
Cyclic voltammograms and bulk electrolysis experiments were
4.4.1 [aqdH2]. Crystals were grown by slow evaporation of
carried out using a Pine Wavenow potentiostat. All CV experi- a solution of aqdH2 in Et2O–hexanes. A single, light yellow
ments were performed in a 0.1 M n-tetrabutylammonium hexa- crystal was used for data collection, with approximate dimen-
fluorophosphate (NBu4PF6)–N,N-dimethylformamide electrolyte sions 0.114 × 0.106 × 0.094 mm. The data were collected at
solution equipped with a glassy carbon button working 25 °C on a Rigaku SCX-Mini diffractometer using a graphite
electrode, a platinum wire counter-electrode, and a silver wire monochromator with MoKα radiation (λ = 0.71073 Å). A total
quasi-reference electrode. The resulting potentials are reported of 920 frames of data were collected using ω and ϕ-scans with
with respect to the ferrocene/ferrocenium redox couple (these a scan range of 1.0° and a counting time of 60.0 seconds per
values can be calculated with respect to SCE by the addition of frame. Data reduction was performed using Rigaku Americas
0.5 V). The bulk electrolysis measurements were carried out in Corporations Crystal Clear version 1.40.60 Details of crystal
a custom H-cell with two frits separating the working and data, data collection and structure refinement are listed in
counter electrode compartments (V ≈ 50 mL). Each cell was Table 1. The oxime bound protons were observed in the elec-
equipped with a glassy carbon rod (Tokai Carbon) as the tron difference map and modeled by a freely rotating riding
Table 1 Crystal data and refinement parameters for the ligand (aqdH2), the parent cobaloxime (1), and the reduced complexes (2–3)
1
2
3
4
[Na(12-crown-4)2]-
[Co(aqdBF2)2(MeCN)2]· [Co(aqdBF2)2(MeCN)]·
[K(dibenzo(18-crown-6)]2
[Co(aqdBF2)2]·2MeCN·Et2O 2[Co(aqdBF2)2]·2THF
[K(18-crown-6) ]-
aqdH2
MeCN
2(MeCN)
Formula
FW
Color
Habit
Size (mm)
T (K)
C12H8N2O2
212.20
Yellow
C32H24B2CoF4N8O4
741.14
Red
C46H53B2CoF4N7NaO12 C72H76B2CoF4K2N6O17
C64H92B2CoF4K2N4O20
1472.16
Deep blue
Block
0.383 × 0.263 × 0.191
150(2)
1075.49
Violet
Shard
1532.13
Deep blue
Shard
Block
Parallelpiped
0.182 × 0.114 × 0.094 0.412 × 0.159 × 0.116
298(2)
0.371 × 0.224 × 0.170
150(2)
0.435 × 0.374 × 0.275
150(2)
173(2)
λ (Å)
0.71073 (MoKα)
Monoclinic
P21/c
7.844(3)
7.8009(29)
15.3089(59)
90
0.71073 (MoKα)
Monoclinic
C2/m
8.9648(3)
18.8380(6)
9.4411(3)
90
1.54184 (CuKα)
Monoclinic
P21/c
12.22900(10)
15.5856(2)
13.20690(10)
90
1.54184 (CuKα)
Monoclinic
P21/n
15.568(6)
20.730(8)
22.817(9)
90
1.54184 (CuKα)
Monoclinic
P21/c
12.9854(6)
19.0412(9)
14.1063(8)
90
Lattice
space group
a (Å)
b (Å)
c (Å)
α (Å)
β (Å)
92.7816(102)
90
935.68(62)
2
1.506
0.106
1.105
R1 = 0.0415
wR2 = 0.0905
99.3490(10)
90
1573.22(9)
3
1.565
0.623
98.8380(10)
90
2487.30(4)
4
1.436
3.509
95.249(7)
90
7333(5)
93.331(5)
90
3482.0(3)
6
1.404
3.725
γ (Å)
V (Å3)
Z
5
dcalc (g cm−3
)
1.388
3.546
1.045
R1 = 0.0533
wR2 = 0.1422
R1 = 0.0650
wR2 = 0.1519
μ (mm−1
)
GOF on F2
R Indices
[I > 2σ(I)]
1.202
0.933
1.042
R1 = 0.0274
wR2 = 0.0814
R1 = 0.0338
wR2 = 0.0961
R1 = 0.0447
wR2 = 0.1223
R1 = 0.0480
wR2 = 0.1263
R1 = 0.0547
wR2 = 0.1493
R1 = 0.0696
wR2 = 0.1637
R Indices all data R1 = 0.0510
wR2 = 0.0956
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
model on the oxime oxygens. They were refined isotropically
without restraints.
4.4.5 [K(18-crown-6)]2[Co(aqdBF2)2] (3′). Crystals of 3′ were
grown at reduced temperature (−20 °C) from a saturated THF
4.4.2 [Co(aqdBF2)2(MeCN)2]MeCN (1). Crystals were grown solution. A blue shard was used for data collection, with
by layering acetonitrile over a saturated solution of 1 in THF. A approximate dimensions 0.383 × 0.263 × 0.191 mm. The data
whole parallel piped crystal was used for data collection, with were collected at 133 K on a Agilent Technologies SuperNova
approximate dimensions 0.401 × 0.164 × 0.126 mm. The data diffractometer using an Atlas S2 Detector and CuKα radiation
were collected at −120 °C on a Nonius Kappa CCD diffract- (λ = 1.54184 Å). Reduced temperatures were maintained by use
ometer using a Bruker AXS Apex II detector and a graphite of an Oxford Cryosystems 700 low-temperature device. A total
monochromator with MoKα radiation (λ = 0.71073 Å). Reduced of 682 frames of data were collected using ω-scans with a scan
temperatures were maintained by use of an Oxford Cryo- range of 1.0° and a counting time of 1.0–15 seconds per frame.
systems 600 low-temperature device. A total of 1048 frames of Data reduction was performed using CrysAlisPro.62 Details of
data were collected using ω and ϕ-scans with a scan range of crystal data, data collection and structure refinement are listed
0.8° and a counting time of 46.24 seconds per frame. Data in Table 1. All non-hydrogen atoms were refined aniso-
reduction was performed using SAINT V8.27B.61 Details of tropically and without restraint.
crystal data, data collection and structure refinement are listed
in Table 1. The hydrogen atoms on C8 were observed in a ΔF
4.5 Physical measurements
map and refined with isotropic displacement parameters, with
EPR spectra were obtained on a Bruker Biospin EMXplus 114
the hydrogen sitting on the mirror plane assigned a site occu-
X-band spectrometer with a liquid nitrogen variable tempera-
ture cryostat. Infrared spectra were recorded using a Bruker
pancy factor of 0.5. All other hydrogen atoms were affixed to
the carbons with an appropriate riding model.
1
Alpha spectrometer equipped with a diamond ATR crystal. H
4.4.3 [Na(12-crown-4)2][Co(aqdBF2)2(MeCN)] (2). Crystals
NMR spectra were collected on a Varian DirecDrive 400 MHz
of 2 were grown by vapor diffusion of Et2O into a saturated
spectrometer. UV/vis absorption spectra were obtained at
solution of the above complex in MeCN. A dark purple frag-
298 K using an Ocean Optics fiber optic system equipped with
ment was broken away from a large crystal and used for data
a low intensity PX-2 pulsed xenon lamp and a USB2000-XR1-ES
detector; samples were prepared as 0.1 mM solutions of the
complexes in THF in 1 cm quartz cuvettes. Solid state mag-
netic susceptibilities were measured at 298 K with a Johnson
collection, with approximate dimensions 0.371 × 0.224 ×
0.170 mm. The data were collected at −123 °C on a Agilent
Technologies SuperNova diffractometer using an Atlas S2
Detector and CuKα radiation (λ = 1.54184 Å). Reduced tempera-
Matthey Mark I instrument.
tures were maintained by use of an Oxford Cryosystems 700
low-temperature device. A total of 1053 frames of data were col-
lected using ω-scans with a scan range of 1.0° and a counting 4.6 Gas chromatography product analysis
time of 1.0–4.5 seconds per frame. Data reduction was per-
Gas samples (5 mL) were drawn from the bulk electrolysis cell
by means of a Vici gas tight syringe (Cat # 050055). The
formed using CrysAlisPro.62 Details of crystal data collection
and structure refinement are listed in Table 1. One of the
samples were injected into a Shimadzu GC-2014 Fuel Cell Ana-
crown-ether molecules was refined without bonds to symmetry
lyzer gas chromatograph, with an argon gas mobile phase. The
products were analyzed on a dual sensor system equipped with
generated atoms via a PART-1 instruction card. The alternate
crown was modelled with positional disorder in the ethylene
bridging units. All non-hydrogen atoms were refined aniso-
tropically, solvent molecules and disordered crown ether
atoms were refined with UDP, but not positional restraint.
4.4.4 [K(dibenzo-18-crown-6)]2[Co(aqdBF2)2] (3). Crystals
a
thermal conductivity detector and a flame ionization
detector.
4.7 DFT calculations
of 3 were grown by vapor diffusion of Et2O into a saturated Geometry optimizations were performed on 1, 2, and 3 using
solution of the above complex in MeCN. A dark blue shard was the coordinates obtained from the crystal structure as starting
used for data collection, with approximate dimensions 0.435 × point. A Hessian calculation that was performed on the opti-
0.374 × 0.275 mm. The data were collected at −123 °C on a mized structures showed no imaginary frequencies. The Firefly
Agilent Technologies SuperNova diffractometer using an Atlas software package63 was utilized in conjunction with the B3P86
S2 Detector and CuKα radiation (λ = 1.54184 Å). Reduced temp- functional and the 6-31G(d,p) basis set with diffuse functions
eratures were maintained by use of an Oxford Cryosystems 700 on heteroatoms and the cobalt center. A larger basis set was
low-temperature device. A total of 1031 frames of data were col- used to perform energy calculations, consisting of the Wach-
lected using ω-scans with a scan range of 1.0° and a counting ters basis set for the cobalt center and the TZVP basis set of
time of 1.0–20 seconds per frame. Data reduction was per- Barbieri et al. for the remaining atoms.64,65 Orbitals and spin
formed using CrysAlisPro.62 Details of crystal data, data collec- density plots were visualized and rendered using Chemcraft.66
tion and structure refinement are listed in Table 1. All non- Initially, spin density calculations on the doubly reduced
hydrogen atoms were refined anisotropically and without species 3 were performed with no restrictions, however, this
restraint with the exception of a rigid bond (RIGU) restraint on lead to spin contamination (S2 = 1.038). Thus, the results dis-
the solvent ether molecule.
played are those of a restricted open shell calculation.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.

View Article Online
Paper
Dalton Transactions
21 Z. Ji, M. He, Z. Huang, U. Ozkan and Y. Wu, J. Am. Chem.
Soc., 2013, 135, 11696–11699.
Acknowledgements
The authors thank the Robert A. Welch Foundation (F-0003, 22 E. S. Andreiadis, P.-A. Jacques, P. D. Tran, A. Leyris,
AHC; F-1822, MJR), ACS Petroleum Research Fund (PRF 53542-
DN13), UT Austin College of Natural Sciences, and the Office
of Naval Research (N00014-13-10530) for financial support. We
M. Chavarot-Kerlidou, B. Jousselme, M. Matheron, J. Pcaut,
S. Palacin, M. Fontecave and V. Artero, Nat. Chem., 2013, 5,
48–53.
also thank Dr Vincent Lynch for assistance with crystallo- 23 A. Krawicz, J. Yang, E. Anzenberg, J. Yano, I. D. Sharp and
graphy, and Mr Mike Ronalter and Mr Adam Kennedy for
expertise in custom glassware fabrication.
G. F. Moore, J. Am. Chem. Soc., 2013, 135, 11861–11868.
24 A. Krawicz, D. Cedeno and G. F. Moore, Phys. Chem. Chem.
Phys., 2014, 16, 15818–15824.
25 A. Fihri, V. Artero, A. Pereira and M. Fontecave, Dalton
Trans., 2008, 5567–5569.
26 P. Zhang, M. Wang, C. Li, X. Li, J. Dong and L. Sun, Chem.
Commun., 2009, 8806–8808.
27 T. M. McCormick, B. D. Calitree, A. Orchard, N. D. Kraut,
F. V. Bright, M. R. Detty and R. Eisenberg, J. Am. Chem.
Soc., 2010, 132, 15480–15483.
Notes and references
1 E. E. Benson, C. P. Kubiak, A. J. Sathrum and J. M. Smieja,
Chem. Soc. Rev., 2008, 38, 89–99.
2 A. J. Morris, G. J. Meyer and E. Fujita, Acc. Chem. Res., 2009,
42, 1983–1994.
3 M. Rakowski Dubois and D. L. Dubois, Acc. Chem. Res., 28 G.-G. Luo, K. Fang, J.-H. Wu, J.-C. Dai and Q.-H. Zhao,
2009, 42, 1974–1982. Phys. Chem. Chem. Phys., 2014, 16, 23884–23894.
4 H. Dau, C. Limberg, T. Reier, M. Risch, S. Roggan and 29 B. D. Stubbert, J. C. Peters and H. B. Gray, J. Am. Chem.
P. Strasser, ChemCatChem, 2010, 2, 724–761. Soc., 2011, 133, 18070–18073.
5 P. Du and R. Eisenberg, Energy Environ. Sci., 2012, 5, 6012– 30 C. C. L. McCrory, C. Uyeda and J. C. Peters, J. Am. Chem.
6021. Soc., 2012, 134, 3164–3170.
6 V. S. Thoi, Y. Sun, J. R. Long and C. J. Chang, Chem. Soc. 31 X. Yin, C. Liu, S. Zhuo, Y. Xu and B. Zhang, Dalton Trans.,
Rev., 2013, 42, 2388–2400. 2014, 44, 1526–1529.
7 J. L. Dempsey, B. S. Brunschwig, J. R. Winkler and 32 S. Troppmann and B. Knig, Chem. – Eur. J., 2014, 20,
H. B. Gray, Acc. Chem. Res., 2009, 42, 1995–2004. 14570–14574.
8 V. Artero, M. Chavarot-Kerlidou and M. Fontecave, Angew. 33 M. Bacchi, G. Berggren, J. Niklas, E. Veinberg, M. W. Mara,
Chem., Int. Ed., 2011, 50, 7238–7266.
9 R. Dreos, S. Geremia, G. Nardin, L. Randaccio, G. Tauzher
and S. Vuano, Inorg. Chim. Acta, 1998, 272, 74–79.
M. L. Shelby, O. G. Poluektov, L. X. Chen, D. M. Tiede,
C. Cavazza, M. J. Field, M. Fontecave and V. Artero, Inorg.
Chem., 2014, 53, 8071–8082.
10 X. Hu, B. M. Cossairt, B. S. Brunschwig, N. S. Lewis and 34 C. N. Valdez, J. L. Dempsey, B. S. Brunschwig, J. R. Winkler
J. C. Peters, Chem. Commun., 2005, 4723–4725.
11 M. Razavet, V. Artero and M. Fontecave, Inorg. Chem., 2005,
44, 4786–4795.
12 X. Hu, B. S. Brunschwig and J. C. Peters, J. Am. Chem. Soc.,
2007, 129, 8988–8998.
and H. B. Gray, Proc. Natl. Acad. Sci. U. S. A., 2012, 109,
15589–15593.
35 P. Kelley, M. W. Day and T. Agapie, Eur. J. Inorg. Chem.,
2013, 2013, 3840–3845.
36 S. M. Laga, J. D. Blakemore, L. M. Henling,
B. S. Brunschwig and H. B. Gray, Inorg. Chem., 2014, 53,
12668–12670.
13 C. Baffert, V. Artero and M. Fontecave, Inorg. Chem., 2007,
46, 1817–1824.
14 B. H. Solis and S. Hammes-Schiffer, Inorg. Chem., 2011, 50, 37 G. N. Schrauzer and R. J. Windgassen, Chem. Ber., 1966, 99,
11252–11262. 602–610.
15 B. H. Solis, Y. Yu and S. Hammes-Schiffer, Inorg. Chem., 38 A. Bakac and J. H. Espenson, J. Am. Chem. Soc., 1984, 106,
2013, 52, 6994–6999. 5197–5202.
16 A. Bhattacharjee, M. Chavarot-Kerlidou, J. L. Dempsey, 39 M. J. Rose, H. B. Gray and J. R. Winkler, J. Am. Chem. Soc.,
H. B. Gray, E. Fujita, J. T. Muckerman, M. Fontecave, 2012, 134, 8310–8313.
V. Artero, G. M. Arantes and M. J. Field, ChemPhysChem, 40 R. van Asselt and C. J. Elsevier, Tetrahedron, 1994, 50, 323–
2014, 15, 2951–2958. 334.
17 J. Niklas, K. L. Mardis, R. R. Rakhimov, K. L. Mulfort, 41 I. L. Fedushkin, A. A. Skatova, V. A. Chudakova and
D. M. Tiede and O. G. Poluektov, J. Phys. Chem. B, 2012,
116, 2943–2957.
18 J. L. Dempsey, J. R. Winkler and H. B. Gray, J. Am. Chem.
Soc., 2010, 132, 1060–1065.
19 J. L. Dempsey, J. R. Winkler and H. B. Gray, J. Am. Chem.
Soc., 2010, 132, 16774–16776.
20 L. A. Berben and J. C. Peters, Chem. Commun., 2010, 46,
398–400.
G. K. Fukin, Angew. Chem., Int. Ed., 2003, 42, 3294–3298.
42 K. Vasudevan and A. H. Cowley, Chem. Commun., 2007,
3464–3466.
43 K. V. Vasudevan, M. Findlater, I. Vargas-Baca and
A. H. Cowley, J. Am. Chem. Soc., 2012, 134, 176–178.
44 I. L. Fedushkin, O. V. Maslova, A. G. Morozov, S. Dechert,
S. Demeshko and F. Meyer, Angew. Chem., Int. Ed., 2012, 51,
10584–10587.
Dalton Trans.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Dalton Transactions
Paper
45 P. Mondal, H. Agarwala, R. D. Jana, S. Plebst, A. Grupp, 57 A. J. C. Wilson, International Tables for X-ray Crystallogra-
F. Ehret, S. M. Mobin, W. Kaim and G. K. Lahiri, Inorg.
Chem., 2014, 53, 7389–7403.
46 N. J. Hill, I. Vargas-Baca and A. H. Cowley, Dalton Trans.,
2008, 240–253.
47 N. G. Connelly and W. E. Geiger, Chem. Rev., 1996, 96, 877–
910.
48 V. Fourmond, P.-A. Jacques, M. Fontecave and V. Artero,
Inorg. Chem., 2010, 49, 10338–10347.
phy, Kluwer Academic Press, Boston, 1992, vol. C.
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ꢀ
ꢀ
2
2 2
ꢀ
ꢀ
Σwð Foj ꢀ Fcj Þ
58 RwðF2Þ ¼
, where w is the weight given
4
ΣwðjFojÞ
each reflection. RðFÞ ¼
ΣðjFojꢀjFcjÞ
for reflections with
ΣjFoj
ꢀ
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ꢀ
2
2 2
ꢀ
ꢀ
Σwð Foj ꢀ Fcj Þ
ðn ꢀ pÞ
Fo > 4ðσðFoÞÞ. GOF; S ¼
, where n is the
49 B. D. McCarthy, D. J. Martin, E. S. Rountree,
A. C. Ullman and J. L. Dempsey, Inorg. Chem., 2014, 53,
8350–8361.
number of reflections and p is the number of refined
parameters.
59 C. F. Macrae, I. J. Bruno, J. A. Chisholm, P. R. Edgington,
P. McCabe, E. Pidcock, L. Rodriguez-Monge, R. Taylor,
J. van de Streek and P. A. Wood, J. Appl. Crystallogr., 2008,
41, 466–470.
60 CrystalClear-SM v. 1.4.0 SP1, Rigaku Americas Corporation,
1998–2008.
50 E. Anxolabhre-Mallart, C. Costentin, M. Fournier,
S. Nowak, M. Robert and J.-M. Savant, J. Am. Chem. Soc.,
2012, 134, 6104–6107.
51 E. Anxolabre-Mallart, C. Costentin, M. Fournier and
M. Robert, J. Phys. Chem. C, 2014, 118, 13377–
13381.
61 SAINT v. 8.27B, Bruker AXS Inc, Madison, WI, 2012.
62 CrysAlisPro. Agilent Technologies, Agilent Technologies UK
Ltd., Oxford, UK, 2013, SuperNova CCD System, CrysAlisPro
Software System, 1.171.37.31.
63 A. Granovsky, Firefly v. 8.0.0, http://classic.chem.msu.su.
64 A. J. H. Wachters, J. Chem. Phys., 1970, 52, 1033–1036.
65 P. L. Barbieri, P. A. Fantin and F. E. Jorge, Mol. Phys., 2006,
104, 2945–2954.
52 M. Satake, J. Miura, S. Usami and B. K. Puri, Analyst, 1989,
114, 813–818.
53 M. C. Burla, R. Caliandro, M. Camalli, B. Carrozzini,
G. L. Cascarano, L. De Caro, C. Giacovazzo, G. Polidori and
R. Spagna, J. Appl. Crystallogr., 2005, 38, 381–388.
54 G. M. Sheldrick, Acta Crystallogr., Sect. A: Fundam. Crystal-
logr., 2008, 64, 112–122.
55 A. L. Spek, Acta Crystallogr., Sect. D: Biol. Crystallogr., 2009,
65, 148–155.
56 L. J. Farrugia, J. Appl. Crystallogr., 2012, 45, 849–854.
66 G. Andrienko, Chemcraft v. 1.8, http://www.chemcraftprog.
com.
This journal is © The Royal Society of Chemistry 2015
Dalton Trans.