Visualizing diastereomeric interactions of chiral amine-chiral copper salen adducts by EPR spectroscopy and DFT

Article abstract of DOI:10.1021/ic200113u

Single enantiomers of R/S-methylbenzylamine (MBA) were found to selectively form adducts with two chiral Cu-salen complexes, [Cu^II(1)] (H ₂1 = N,N′-bis(3,5-ditert-butylsalicylidene)-1,2- diaminocyclohexane) and [Cu^II(2)] (H₂2 = N,N′-bis-salicylidene-1,2-cyclohexanediamino). The axial g/A spin Hamiltonian parameters of the Cu-MBA adducts were typical of 5-coordinate species. Enantiomer discrimination in the MBA binding was directly evidenced by W-band CW EPR, revealing an 86 ± 5% preference for formation of the R,R-[Cu(1)] + S-MBA adducts compared to R,R-[Cu(1)] + R-MBA; this was reduced to a 57 ± 5% preference for R,R-[Cu(2)] + S-MBA following removal of the tert-butyl groups. The structure of these diastereomeric adducts was further probed by different hyperfine techniques (ENDOR and HYSCORE), although no structural differences were detected between these adducts using these techniques. The diastereomeric adducts were found to possess lower symmetry, as evidenced by rhombic g tensors and inequivalent H^imine couplings. This was caused by the selective binding mode of MBA onto one side of the chiral Cu^II complex. DFT calculations were performed on the R,R-[Cu(1)] + S-MBA and R,R-[Cu(1)] + R-MBA adducts. A distinct difference in orientation and binding mode of the MBA was identified in both adducts, confirming the experimental results. The preferred heterochiral R,R-[Cu(1)] + S-MBA adduct was found to be 5 kJ mol^-1 lower in energy compared to the homochiral adduct. A delicate balance of steric repulsion between the α-proton (attached to the asymmetric carbon atom) of MBA and the methine proton (attached to the asymmetric carbon atom) of [Cu(1)] was crucial in the stereoselective binding.

Full text of DOI:10.1021/ic200113u

ARTICLE
pubs.acs.org/IC
Visualizing Diastereomeric Interactions of Chiral AmineꢀChiral
Copper Salen Adducts by EPR Spectroscopy and DFT
Damien M. Murphy,*,† Ignacio Caretti,‡ Emma Carter,† Ian A. Fallis,*,† Marcus C. G€obel,† James Landon,†
Sabine Van Doorslaer,*,‡ and David J. Willock†
†School of Chemistry, Cardiff University, Main Building, Park Place, Cardiff CF10 3AT, United Kingdom
‡SIBAC Laboratory, Department of Physics, University of Antwerp, Universiteitsplein 1, B-2610 Wilrijk, Belgium
S Supporting Information
b
ABSTRACT: Single enantiomers of R/S-methylbenzylamine (MBA) were found to
selectively form adducts with two chiral Cuꢀsalen complexes, [CuII(1)] (H21 = N,
N0-bis(3,5-ditert-butylsalicylidene)-1,2-diaminocyclohexane) and [CuII(2)] (H22 =
N,N0-bis-salicylidene-1,2-cyclohexanediamino). The axial g/A spin Hamiltonian
parameters of the CuꢀMBA adducts were typical of 5-coordinate species. En-
antiomer discrimination in the MBA binding was directly evidenced by W-band CW
EPR, revealing an 86 ( 5% preference for formation of the R,R-[Cu(1)] þ S-MBA
adducts compared to R,R-[Cu(1)] þ R-MBA; this was reduced to a 57 ( 5%
preference for R,R-[Cu(2)] þ S-MBA following removal of the tert-butyl groups.
The structure of these diastereomeric adducts was further probed by different
hyperfine techniques (ENDOR and HYSCORE), although no structural differences
were detected between these adducts using these techniques. The diastereomeric
adducts were found to possess lower symmetry, as evidenced by rhombic g tensors
and inequivalent Himine couplings. This was caused by the selective binding mode of MBA onto one side of the chiral CuII complex.
DFT calculations were performed on the R,R-[Cu(1)] þ S-MBA and R,R-[Cu(1)] þ R-MBA adducts. A distinct difference in
orientation and binding mode of the MBA was identified in both adducts, confirming the experimental results. The preferred
heterochiral R,R-[Cu(1)] þ S-MBA adduct was found to be 5 kJ molꢀ1 lower in energy compared to the homochiral adduct. A
delicate balance of steric repulsion between the R-proton (attached to the asymmetric carbon atom) of MBA and the methine
proton (attached to the asymmetric carbon atom) of [Cu(1)] was crucial in the stereoselective binding.
’ INTRODUCTION
for asymmetric catalysis was performed by Jacobsen and
Katsuki.5,9 Chromium and cobalt complexes of the same N,N0-
bis(3,5-di-tert-butylsalicylidene)-1,2-cyclohexanediamino ligand
(1) (Scheme 1) were also found to catalyze the highly enantio-
selective ring-opening reactions of epoxides.10
Despite the phenomenal success of these catalysts, a great deal
of research has been devoted to understanding their mode of
operation. This is particularly true for the epoxidation reaction,
where many key aspects of the reaction are still not fully under-
stood (i.e., nature of oxygen-transferring species, mechanism of
the oxygen transfer step, nature of highly efficient stereochemical
communication between catalyst and substrate).7 Several ap-
proaches, including computational and spectroscopic methods,
have been applied to explore how the steric properties of
the ligand contribute to the asymmetric chiral induction mechan-
ism and to verify the precise transition state of the asymmetric
reaction.11
In recent years transition-metal complexes have been exten-
sively used in asymmetric catalytic transformations of racemic,
prochiral, and optical substrates to yield single-enantiomer
compounds.1 The importance of this field was recognized
10 years ago, when Knowles, Noyori, and Sharpless received
the Nobel Prize in chemistry for their work on chirally catalyzed
hydrogenation/oxidation reactions.2 To optimize efficiency,
these catalysts should strictly differentiate enantiotopic groups
or faces of prochiral molecules.3 Optimization can be modulated
based on the choice of metal (Cu, Mn, Co, Fe, Cr, etc.) or the
choice of chiral organic ligands coordinated to the metal center.
Ligands with central chirality, axial chirality, and planar chirality
can thus be easily tailored in order to fine tune the desired
enantioselectivities.4
One very versatile class of organic ligand for homogeneous
asymmetric catalysis is the salen ligand [1,6-bis(2-hydroxy-
phenyl)-2,5-diazahexa-1,5-diene].5 Much of this versatility origi-
nates from the structural rigidity of the ligand,6 which is remin-
iscent of the porphyrin framework in heme-based oxidative
enzymes;7 the ligand can thus easily stabilize metal ions in an
asymmetric environment yet is easy to synthesize and derivatize
relative to porphyrins.8 Early pioneering work on these ligands
One very versatile but often under exploited spectroscopic
tool for studying homogeneous catalytic reactions is electron
paramagnetic resonance (EPR) and the related hyperfine
Received: January 18, 2011
Published: June 27, 2011
r
2011 American Chemical Society
6944
dx.doi.org/10.1021/ic200113u Inorg. Chem. 2011, 50, 6944–6955
|

Inorganic Chemistry
ARTICLE
Scheme 1. Structure of the N,N0-Bis(3,5-di-tert-butylsalicy-
lidene)-1,2-cyclohexanediamino Metal Complex, N,N0-Bis-
(3,5-di-tert-butylsalicylidene)-1,2-cyclohexanediamino
Copper(II) ([Cu(1)], and the N,N0-Bis-salicylidene-1,2-
cyclohexanediamino Copper(II) ([Cu(2)]) complexa
mode of stereoselective binding in asymmetric metal complexes
with organic substrates bearing phenyl groups. While advanced
EPR techniques have been widely used to examine the nature of
the copper binding sites in various proteins, metalloenzymes,23
or even relatively simple organometallic complexes,24 they
have not been widely employed to study asymmetric complexes
of relevance to catalysis. In this work we will show how
high-frequency EPR can be used to examine the binding site in
an asymmetric metal complex bearing a coordinated asymmetric
substrate. Using W-band EPR, we will show the extent of enan-
tiodiscrimintaion in [Cu(1)] versus [Cu(2)]. ENDOR and
HYSCORE have also been used to investigate the structure
of the adducts, although in this case these techniques were unable
to detect any differences between the diastereomeric adducts.
The origins of this diastereoselective binding is explained
by DFT.
’ EXPERIMENTAL SECTION
Synthesis. All reagents and solvents (Aldrich) were used as
received. Ligands H21 and H22 were prepared as described previously.20
Complexes [Cu(1)] and [Cu(2)] (Scheme 1) were prepared by the
methods of Bunce et al.25a and Bernado et al.,25b respectively. Complexes
were purified by filtration through a short pad of silica to afford chroma-
tographically homogeneous samples. d2-MBA was prepared by washing
a CH2Cl2 (50 mL) solution of MBA (5 g) with several portions of D2O
(3 ꢁ 30 mL). The organic layer was dried (MgSO4), and the solvent was
removed in vacuo. The residue was distilled to afford d2-MBA in quan-
titative yield. The degree of deuteration was estimated to be ∼90%
(analysis by electrospray mass spectrometry).
X-Band Pulsed EPR. The X-band pulsed EPR and pulsed ENDOR
experiments were performed with an ELEXSYS Bruker spectrometer
(mw frequency 9.73 GHz) equipped with a liquid-helium cryostat from
Oxford Inc. All experiments were recorded at 10 K with a repetition rate
of 1kHz. The magnetic field was measured with a Bruker ER035 M NMR
Gaussmeter. HYSCORE experiments27 were carried out using the pulse
sequence π/2ꢀτꢀπ/2ꢀt1ꢀπꢀt2ꢀπ/2ꢀτꢀecho, with pulse lengths
of tπ/2 = tπ = 16 ns. Times t1 and t2 were varied from 96 to 5680 ns in
steps of 16 ns. An eight-step phase cycle was used to eliminate unwanted
echoes. The individual time traces were baseline corrected with a third-
order polynomial, apodized with a Hamming window, and zero filled.
After 2D Fourier transformation, the absolute-value spectrum was
calculated. The HYSCORE spectra were simulated using a GAMMA-
based program developed at the ETH Zurich.28 The same sets of τ values
as in the experiments were taken. Davies-ENDOR experiments29 were
carried out using the following pulse sequence: πꢀTꢀπ/2ꢀτꢀ
πꢀτꢀecho. The experiments were done with mw pulse lengths of
tπ = 256 ns, tπ/2 = 128 ns, and an interpulse time τ of 800 ns. An rf
τ pulse of variable frequency and a length of 18 μs was applied during
time T of 20 μs.
W-Band CW and Pulsed EPR. The W-band EPR measurements
were performed on a Bruker Elexsys E680 spectrometer equipped with a
continuous gas-flow cryostat (Oxford Instruments). The spectra were
recorded using microwaves of ∼94 GHz frequency and 0.044 mW
power at a fixed temperature of 100 K. The magnetic field was swept in
the 2950ꢀ3350 mT range with a modulation frequency and amplitude
of 100 kHz and 0.5mT, respectively. The W-band ELDOR (electronꢀ
electron double resonance)-detected NMR experiments30 were per-
formed using the pulse sequence (HTA)mw2ꢀt1ꢀ(π/2)mw1ꢀτꢀ
(π)mw1ꢀτꢀecho with (π/2)mw1 and (π)mw1 pulses of 220 and
440 ns, a 50 μs high-turning angle (HTA)mx2 pulse with mw frequency
2, and interpulse delay times t1= 1 μs and τ2 = 2 μs. The experiments
were performed at 5 K with a repetition rate of 667 Hz.
a The structure of the chiral methylbenzylamine (MBA) is also shown.
Chiral centers are labeled with an asterisk (*).
techniques of ENDOR (electron nuclear double resonance),
ESEEM (electron spin echo envelope modulation), and HY-
SCORE (hyperfine sublevel correlation) spectroscopy.12,13 For
example, in the specific case of metalꢀsalen-based catalysts,
active MnVdO species in the [MnCl(1)] complex have been
investigated by EPR.14ꢀ17 The Cr spin states in [CrCl(1)], of
relevance to catalytic epoxidation, have also been studied.18
Recently, we investigated the electronic structure of acetic-
acid-activated [Co(1)], of relevance to hydrolytic kinetic reso-
lution,19 and identified an unusual coordinated phenoxyl radical,
[CoIII(1•)(OAc)n](OAc)m, among several paramagnetic centers
in the activated catalyst.19
In most of these investigations, the emphasis of the work was
devoted to the electronic properties of the central metal15ꢀ18 or
ligand.19 However, as reported by Movassaghi and Jacobsen,20
homogeneous asymmetric metal catalysts can behave like en-
zymes by exploiting hydrogen bonding between the active site
and substrate together with nonbonded dipoleꢀdipole, electro-
static, and steric interactions to orient the substrate and stabilize
the transition state, leading to high levels of stereoselectivity. We
have already shown how some of these weak outer-sphere forces
can be studied by EPR, ENDOR, HYSCORE, and DFT in the
related chiral [VO(1)] undergoing interaction with chiral ep-
oxides; in particular, we showed how weak electrostatic forces,
steric interactions, and H bonds affect the manner in which the
chiral epoxides bind to the chiral [VO(1)] complexes.21
In this work, we will examine how the asymmetric copper
complexes [Cu(1,2)] (Scheme 1) can differentiate and discri-
minate between single enantiomers of a chiral amine (R- and
S-methylbenzylamine). Our choice of chiral ligands 1 and 2 was
motivated by a desire to establish the possible role of πꢀπ
interactions, H bonds, or simple steric repulsion in controlling
the outcome of these diastereomeric interactions. For example, it
is well known that πꢀπ interactions control the conformational
preferences and binding in macrocycles, complexation in many
hostꢀguest systems, porphyrin aggregation, and the structure of
DNA.22 It is not clear if these and other weak forces control the
6945
dx.doi.org/10.1021/ic200113u |Inorg. Chem. 2011, 50, 6944–6955

Inorganic Chemistry
ARTICLE
Computational Details. The interaction of R- and S-MBA with
R,R-[Cu(1)] was studied computationally using a mixed quantum
mechanics/molecular mechanics (QMMM) approach through the
ONIOM method31 within the Gaussian03 package.32 The metal com-
plex including the metal center, phenyl, and methine groups along with
the MBA were treated at the QM (BHandH) level, while the remaining
aliphatic tert-butyl groups and cyclohexyl backbone were studied at the
MM level (universal force field, UFF).33 This division of the atoms is
illustrated in the Supporting Information (Figure S13) and is chosen to
ensure that the boundary region consists of only σ bonds. Hydrogen link
atoms are used to satisfy the valency of the QM region at this interface.
One aspect of the MBA interaction with [Cu(1)] is the presence of
aromatic groups in both species. Selection of BHandH was based on
recent work comparing a range of functionals for describing benzene,
pyridine, and DNA base dimers.34 This clearly shows that minima on the
potential energy surfaces with the BHandH functional give qualitative
agreement with MP2 structures and can reproduce interaction energies
generally to within (2 kJ molꢀ1. BHandH is a hybrid functional with the
exchange component of the electronꢀelectron interaction represented
using one-half HartreeꢀFock and one-half local spin density ap-
proaches. The correlation energy is drawn from the functional due to
Lee, Yang, and Parr.35 Within the QM region we also employ the
LANL2DZ effective core potential (ECP) and related basis set for Cu,
the 6-31G(d,p) basis for C, O, and H, and 6-31þG(d,p) for N. The
additional diffuse function at N was introduced to improve the descrip-
tion of the MBAꢀ[Cu(1)] interaction involving the coordinating
nitrogen atom.
In order to compute the EPR parameters, spin-unrestricted density
functional computations were performed with the ORCA package36 on
the two most stable geometries of R,R-[Cu(1)] þ S-MBA and the single
most stable geometry of R,R-[Cu(1)] þ R-MBA. The computations
were performed with the B3LYP functional. For calculation of the EPR
parameters, basis sets with significant flexibility in the core region were
used, i.e., the ORCA basis sets CoreProp (CP(III))37 was used for
copper, and a Barone basis set ‘EPRII’38 was taken for the nitrogen and
hydrogen atoms. For the other atoms, a split-valence plus polarization
(SV(P)) basis set was assumed.39 The solvent surrounding was simu-
lated by assuming a dielectric surrounding with the dielectric constant of
MBA (ε = 4.4 at 18 ꢀC) using the conductor-like screening model
(COSMO).40
and heterochiral combinations (i.e., RR in S-MBA and SS in
R-MBA). The preferential formation of one adduct over another
was furtherconfirmed inseparate competitive experiments, where-
by either rac-[Cu(1,2)] was dissolved in rac-MBA (denoted as
a ‘rac-rac’ combination) or alternatively one enantiomer of
[Cu(1,2)] was dissolved in rac-MBA (denoted RR-rac or SS-rac
combinations).
’ CW-EPR SPECTRA (X BAND)
The X-band CW-EPR spectra for [Cu(1)] dissolved in MBA
are shown in Figure 1bꢀf together with the parent spectrum of
rac-[Cu(1)] dissolved in toluene (Figure 1a). Figure 1b shows
the EPR spectrum of the competitive rac-rac combination,
Figure 1c and 1d shows the spectra of the heterochiral pairs,
while Figure 1e and 1f shows the homochiral pairs. The equi-
valent series of spectra for [Cu(2)] are shown in the Supporting
Information (Figure S1).
The EPR spectra of [Cu(1)] and [Cu(2)] measured in
toluene (i.e., in the absence of amine) display axial g and copper
hyperfine (ACu) tensors. This was confirmed at W-band frequen-
cies (Figures 2d and 3d). Further superhyperfine splittings are
also readily observed at X-band (Figure 1) due to the strong
interaction of the unpaired electron with the two equivalent
nitrogen nuclei and two equivalent imine protons of the Schiff-
base ligand (Scheme 1). These proton and nitrogen superhyper-
fine splittings are expected for 4-coordinate Cu(II) complexes.41
The spin Hamiltonian parameters obtained from the simulated
EPR spectra are listed in Table 1. These parameters are in good
agreement with those obtained from a single-crystal EPR study
of the closely related [Cu(salen)] complex.42 EPR spectra for
frozen solutions of [Cu(1)] were also recently reported within
the context of galactose oxidase-inspired complexes.43
The individual spectra of [Cu(1)] in MBA are shown in
Figure 1bꢀf and are obviously different compared to Figure 1a
(of [Cu(1)] in toluene); this is indicative of amine binding to the
copper complex. The most notable changes are the increasing g||
and decreasing A|| values; these trends are consistent with the
weak axial coordination of a fifth ligand (MBA) to the poorly
Lewis acidic [Cu(1)]. The resolution of the superhyperfine
splittings in the g = g^ region is substantially reduced compared
to the spectrum observed in the absence of amine (Figure 1a).
This loss of resolution is largely due to the subtle changes in the
1H and 14N hyperfine couplings arising from the redistribution of
spin density on the nitrogen and imine protons, which generally
accompanies the axial coordination of a fifth ligand to copper
centers.41 The loss in superhyperfine resolution is even more
pronounced for the [Cu(2)] complex (Supporting Information;
Figure S1). This suggests that MBA is weakly coordinated
to [Cu(1,2)] but does not reveal the number of bound MBA
molecules.
To answer this question, a series of dilution measurements was
performed in which the Cu:MBA ratios were systematically
varied from 1 to 20 for a constant concentration of [Cu(1)] or
[Cu(2)]; these spectra were compared to the two limiting situa-
tions of [Cu(1,2)] recorded in neat toluene/dichloromethane or
in excess MBA. The resulting spectra could be treated as a linear
combination of the individual spectra observed in the presence
and absence of MBA (see Supporting Information; Figure S2),
and thus, it was possible to determine the number of coordinated
MBA molecules. For both complexes, this revealed that only one
bulky MBA molecule was bound to the bulky Cu complex. This is
’ RESULTS
The chiral amines (R- and S-MBA) and chiral copper com-
plexes (R,R- and S,S-[Cu(1,2)] can form four different 5-co-
ordinate diastereomeric adducts; namely, R,R-[Cu(1,2)] þ R-
MBA, R,R-[Cu(1,2)] þ S-MBA, S,S-[Cu(1,2)] þ R-MBA, and
S,S-[Cu(1,2)] þ S-MBA (abbreviated hereafter to RR-R, RR-S,
SS-R, and SS-S, see Scheme S1, Supporting Information). The
assumption that only 5-coordinate Cu(II) adducts are formed
was verified in various dilution experiments (see later). Since the
two adducts RR-R and SS-S are enantiomers of each other and
will have the same structure and stability, they are henceforth
referred to as the “homochiral” pairs or adducts. Similarly, the
adducts RR-S and SS-R are enantiomers and will be referred to as
the “heterochiral” pairs or adducts. The homochiral and hetero-
chiral pairs of enantiomers are diastereomers and have different
structures, stabilities, and spectroscopic properties. X- and
W-band CW EPR spectra were therefore recorded for different
combinations of the copper complexes and chiral amines to test
which adducts are preferentially formed. This involved first
measuring the EPR spectra of the individual homochiral combi-
nations (i.e., RR dissolved in R-MBA and SS dissolved in S-MBA)
6946
dx.doi.org/10.1021/ic200113u |Inorg. Chem. 2011, 50, 6944–6955

Inorganic Chemistry
ARTICLE
Figure 1. X-band CW EPR spectra (recorded at 140K) of (a) the noncoordinated [Cu(1)] complex dissolved in toluene together with the different
[Cu(1)] þ MBA combinations of (b) rac-rac, (c) RR-S, (d) SS-R, (e) RR-R, and (f) SS-S.
Figure 3. W-band CW EPR spectra (recorded at 100K) of S,S-[Cu(2)]
dissolved in (a) rac-MBA, (b) S-MBA, (c) R-MBA, and (d) DCM. The
corresponding simulations (a0, b0, c0, and d0) are plotted below each
spectrum; (a0) obtained by addition of experimental spectra c and b in a
ratio of 57:43.
Figure 2. W-band CW EPR spectra (recorded at 100K) of S,S-[Cu(1)]
dissolved in (a) rac-MBA, (b) S-MBA, (c) R-MBA, and (d) toluene. The
corresponding simulations (a0, b0, c0, and d0) are plotted below each
spectrum; (a0) obtained by addition of experimental spectra c and b in a
ratio of 86:14.
not surprising considering the propensity for Cu(II) complexes
Closer inspection of the spectra in the g = g^ region reveals
to form 5 (i.e., 4 þ 1) coordinate complexes.
some interesting features (Figure 1). The EPR spectra of the
6947
dx.doi.org/10.1021/ic200113u |Inorg. Chem. 2011, 50, 6944–6955

Inorganic Chemistry
ARTICLE
Table 1. Experimental and DFT-Computed Principal g and 63Cu Spin Hamiltonian Parameters for rac-[Cu(1)] and rac-[Cu(2)]
Dissolved in Toluene (tol) and Dichloromethane (dcm)a
solvent/matrix
g3
g2
g1
A3 /MHz
A2 /MHz
A1 /MHz
rac-[Cu(1)] þ tol
2.1937
2.1957
2.2226
2.2235
2.2310
2.2230
2.0408
2.0416
2.0475
2.0436
2.0442
2.0428
2.0408
2.0416
2.0475
2.0532
2.0600
2.0600
ꢀ605 ( 10
ꢀ617 ( 10
ꢀ565 ( 10
ꢀ550 ( 10
ꢀ535 ( 10
ꢀ550 ( 10
ꢀ90 ( 5
ꢀ90 ( 5
ꢀ90 ( 5
ꢀ60 ( 10
ꢀ60 ( 10
ꢀ65 ( 10
ꢀ65 ( 10
rac-[Cu(2)] þ dcm
ꢀ90 ( 5
ꢀ60 ( 10
ꢀ60 ( 10
ꢀ65 ( 10
ꢀ65 ( 10
bR,R-[Cu(1)] þ R-MBA
R,R-[Cu(1)] þ S-MBA
R,R-[Cu(2)] þ R-MBA
R,R-[Cu(2)] þ S-MBA
gz
gy
gx
Az/MHz
Ay /MHz
Ax /MHz
A
B
C
2.1504
2.0532
2.0525
2.0543
2.0380
2.0385
2.0375
ꢀ565.3
ꢀ565.3
ꢀ563.6
ꢀ9.5
ꢀ6.1
ꢀ72.6
ꢀ67.6
ꢀ76.7
2.1492
2.1513
ꢀ12.4
a The DFT-computed principal values correspond to the optimized structures A and B of the R,R-[Cu(1)] þ R-MBA adduct and structure C for the R,
R-[Cu(1)] þ S-MBA adduct (Figure 5). b Only the values for R,R-[Cu(1,2)] þ R/S-MBA are given, since S,S-[Cu(1,2)] þ R/S-MBA gives similar
results; g1,g2 = g^; g3 = g||. A1,A2 = A^; A3 = A||. A^ values obtained from X-band spectra. Error in g values is (0.0005
heterochiral pairs (Figure 1c and 1d) are identical to each other,
as are the homochiral combinations (Figure 1e and 1f). Satisfy-
ingly, a comparison of the heterochiral (Figure 1c and 1d) and
homochiral (Figure 1e and 1f) structures reveals differences in
the spectra, as expected for a set of diastereomers. Measurement
of the rac-rac spectrum (Figure 1b) was undertaken to determine
the extent, if any, of chiral discrimination in the system. Theore-
tically, if both homochiral and heterochiral adducts were equally
formed in a rac-rac solution then, under the magnetically dilute
conditions employed, Figure 1b should represent an equal
summation of Figure 1c and 1d and Figure 1e and 1f. In fact,
Figure 1b is comparatively closer to Figure 1c and 1d compared
to Figure 1e and 1f. This important observation suggests that
there is a preferential formation of the heterochiral adducts in
frozen solution and that [Cu(1)] is capable of discriminating
enatiomers of MBA.
evidence for a different binding mode of the MBA in the homo-
chiral pair (SS-S) than in the heterochiral one (SS-R), where axial
nature is retained, and indicates a lowering of symmetry at the
metal center in the homochiral case. Again, this is evidence for a
discrimination of the MBA enantiomers by the [Cu(1)] com-
plex. The W-band EPR spectrum of S,S-[Cu(1)] dissolved in rac-
MBA (SS-rac) is shown in Figure 2a, with a simulation based
on an 86:14 contribution of SS-R:SS-S (using the least-squares
method). Therefore, the S,S-[Cu(1)] complex preferentially
binds the R-MBA enantiomer with a selectivity of (86 ( 5)%
(i.e., favoring the heterochiral pair). These results quantify the earlier
qualitative trends observed in the X-band EPR spectra (Figure 1).
A similar set of experiments was also conducted using
[Cu(2)]. The W-band CW-EPR spectra of S,S-[Cu(2)] dis-
solved in rac-MBA, R-MBA, S-MBA, and CH2Cl2 (owing to the
poor solubility of [Cu(2)] in toluene) are shown in Figure 3.
Similarly to [Cu(1)], [Cu(2)] exhibits a square-planar geometry
with a gz value 0.002 larger than in [Cu(1)] and where A3 has
increased in absolute value by 10 MHz. The binding of R- or
S-MBA to S,S-[Cu(2)] now results in a pronounced rhombic
distortion (Table 1) as shown by two well-resolved peaks in the
perpendicular region. These extra peaks cannot arise from a
second Cu(II) center, since no extra peaks are detected in the
parallel region, neither at the X band nor at the W band. Spectral
simulations for [Cu(2)]ꢀMBA adducts revealed that the hetero-
chiral adduct (SS-R) is preferred over the homochiral adduct
(SS-S) but with only a 57:43 contribution of SS-R:SS-S, indicating
that the heterochiral pair is only slightly favored at 57((5)%.
This key observation indicates a poorer discrimination of en-
atiomers of MBA by [Cu(2)] compared to that of [Cu(1)].
’ CW-EPR SPECTRA (W BAND)
The high-field EPR spectra of S,S-[Cu(1)] dissolved in rac-,
R-, and S-MBA are shown in Figure 2a, 2b, and 2c, respectively.
Figure 2d corresponds to S,S-[Cu(1)] dissolved in toluene for
comparison. Simulations are depicted below each experimental
spectrum and labeled as Figure 2a0ꢀd0. It should be noted that
the g-strain effect is 10 times larger at W-band frequencies
compared to the X band; as a result, the hyperfine splitting
resolution is lost in the g = g^ region. The narrower line width
of the complex dissolved in toluene (Figure 2d) compared to
MBA (Figure 2aꢀc) points to smaller g and A strain effects in
the noncoordinated [Cu(1)]ꢀtoluene case compared to the
[Cu(1)]ꢀMBA adducts.
As observed at the X band, the W-band EPR spectra in Figure 2
reveal that amine coordination has a considerable effect on
the spin distribution in the copper center. Consequently, the
[Cu(1)]ꢀMBA adducts show an increase of the g values and a
decrease of the copper hyperfine constants (in absolute values)
compared to [Cu(1)] (Table 1). This variation of g and A
further confirms axial coordination of the MBA. Examination of
the spin Hamiltonian parameters of [Cu(1)]ꢀMBA diastereo-
mers in Table 1 reveals a small rhombicity of the g tensor for the
homochiral pair (SS-S) that could not be detected at X-band
frequencies. This deviation of the g tensor from axial symmetry is
’ ENDOR AND HYSCORE SPECTRA
To probe the structure of the diastereomeric MBA adducts
further, ENDOR and HYSCORE were used to examine the
1H and 14N couplings from the ligand and coordinated MBA
molecule (ꢀNH2). Although relevant changes in the hyperfine
parameters of the ligand nuclei can be detected by ENDOR and
HYSCORE, no significant difference between the diastereomeric
adducts could be revealed using these hyperfine techniques. A
summary of the key findings from these hyperfine techniques is
thus presented below.
6948
dx.doi.org/10.1021/ic200113u |Inorg. Chem. 2011, 50, 6944–6955

Inorganic Chemistry
ARTICLE
Figure 4. Q-band 1H CW ENDOR spectra (recorded at 10K) showing the Himine couplings of (a) rac-[Cu(2)] dissolved in d8-toluene/d6-
dichloromethane and (b) SS-R, (c) RR-S, (d) SS-S, and (e) RR-R. Spectra recorded at (A) g = g = g3 and (B) g = g^= (g1 þ g2)/2. For clarity, the central
part of the spectra, containing proton couplings from the remaining ligand nuclei and coordinated MBA, have been removed.
The Q-band ENDOR spectra of [Cu(2)] recorded at field
positions corresponding to g = g^ and g = g3 = g are shown in
Figure 4. The corresponding hyperfine values extracted from the
simulations are given in Table 2. The analogous series of spectra
for [Cu(1)] are shown in the Supporting Information (Figure S5).
In the absence of MBA, the two imine protons appear equivalent
for both [Cu(1)] and [Cu(2)]. This equivalency can be clearly
seen in Figure 4a for the spectrum measured at g = g , where only
one pair of lines is observed. The aiso values and CuꢀHimine
distances of 3.8 Å (estimated from ADip using a simple point-
dipole approximation) are in very good agreement with the single
crystal of [Cu(1,2)]25a,b and the structurally related [Cu(salen)]
complex.42
Upon addition of MBA, the equivalency of the imine protons
is removed. The spectra of the homochiral (RR-R, SS-S) and
heterochiral (RR-S, SS-R) adducts are given in Figure 4bꢀe; two
distinct differences can now be observed. First, a small decrease
in the magnitude of aiso accompanies MBA coordination to
[Cu(1)] and [Cu(2)] (see Supporting Information; Table S1). A
decrease of aiso was also reported for the imine protons upon axial
coordination of a fifth ligand (pyridine) to [Cu(salen)],42 caused
by a loweringofcovalency in the complex. Second, the equivalency
of the two imine protons is removed upon MBA binding and two
distinct protons (labeled HA and HB) can now be distinguished
(Table 2). Again, thisinequivalency can be seen in Figure4bꢀe for
the spectra measured at g = g ; two pairs of lines are now clearly
visible centered on νH. This inequivalency in the ligand imine
protons is due to a lowering in the symmetry of the [Cu(2)] þ
MBA adduct, compared to the parent [Cu(2)] complex. These
results supportthe earlierdirectfindings by W-bandEPR, whereby
a rhombic distortion was found for the [Cu(2)] þ MBA adduct
(Figure 3), indicative of a lower symmetry.
these results support the earlier W-band data, whereby the
rhombic distortion in the [Cu(1)] þ MBA adduct (Figure 2)
was almost negligible compared to [Cu(2)] (Figure 3). We next
examined the couplings to the ligand methine protons in
[Cu(1,2)] and tert-butyl protons in [Cu(1)] by CW and Davies
ENDOR (experimental and simulated spectra shown in the
Supporting Information; Figures S6ꢀS8). The hyperfine cou-
plings (hfc) for these two sets of protons are given in Table 2. It
should be clearly noted that these hyperfine couplings did not
alter dramatically in the presence or absence of MBA. This
suggests little interaction between the bound MBA and these
protons nor any substantial distortions of the [Cu(1,2)] ligand
framework. We also examined the ENDOR spectra arising
from the amine protons (ꢀNH2) of the coordinated MBA
(Supporting Information; Figures S9ꢀS10). The assignment of
the peaks due to the ꢀNH2 group was confirmed by comparison
of the ENDOR spectra recorded for [Cu(1,2)] in the absence of
MBA with those recorded using MBA and d2-MBA (Supporting
Information; Figures S11). The simulated hyperfine couplings
for the amine protons are listed in Table 2. These values are
expected for a weakly coordinated amine. Despite recording the
ENDOR spectra of the amine protons for a set of diastereomeric
adducts (RR þ S versus RR þ S), we could not detect any
significant differences.
Q-band 14N ENDOR spectra of [Cu(1,2)] in toluene and
MBA were also measured (Supporting Information; experimen-
tal and simulated Q-band CW spectra shown in Figures S12 and
S13 and X-band Davies ENDOR is shown in Figure S8). For the
parent complexes dissolved in toluene/dichloromethane, both
the hfc and the nuclear quadrupole tensors of the two nitrogen
nuclei were found to deviate slightly from axial symmetry; their
largest principal axes were approximately directed to the copper
ion. The simulated parameters (Table 3) are very similar to those
reported for [Cu(salen)].42
In the [Cu(1)] þ MBA adduct, the inequivalency in the imine
protons was barely perceptible, manifested only by a broadening of
the proton peaks in the ENDOR simulations, rather than a complete
splitting of the peaks (Supporting Information; Figure S5). Again,
Upon coordination of MBA, a reduction in the 14N hfc is
observed for both [Cu(1,2)] and is analogous with the earlier
6949
dx.doi.org/10.1021/ic200113u |Inorg. Chem. 2011, 50, 6944–6955

Inorganic Chemistry
ARTICLE
Table 2. Continued
observation of a reduction in aiso for the imine protons
(Figure 4); this is again consistent with axial coordination of
MBA. A noted decrease in the asymmetry parameter η of the
bound adduct can also be observed for [Cu(2)], while it is less
pronounced for [Cu(1)]. The 14N ENDOR spectra were also
notably broader (and less intense) for the [Cu(1,2)]-MBA
adducts compared to the parent complexes (Supporting Infor-
mation; Figures S12 and S13). While solvation issues cannot be
ignored completely (i.e., MBA vs toluene/DCM), this broad-
ening of the lines may arise from an inequivalency of the two
ligand nitrogen nuclei in the adducts. These observations corro-
borate the inequivalency observed above for the imine protons.
The hyperfine interaction with the axial amine nitrogen is
expected to be very small and could not be observed in the CW
DFT
A2
experimental
A2
nuclei
A1
A3
A1
A3
amine protons
[Cu(1)]
ꢀ2.52 ꢀ4.87
ꢀ2.62 ꢀ5.20
3.62
3.62
[Cu(2)]
A H (ꢀNH2)
H (ꢀNH2)
ꢀ4.08 ꢀ4.40
ꢀ3.34 ꢀ3.61
ꢀ3.09 ꢀ2.88
ꢀ4.16 ꢀ4.51
ꢀ2.60 ꢀ2.74
ꢀ3.95 ꢀ4.54
ꢀ1.27 ꢀ1.32
ꢀ1.88 ꢀ2.03
ꢀ2.27 ꢀ2.42
7.73
6.25
5.03
7.94
4.50
7.84
2.71
3.90
4.80
B H (ꢀNH2)
H (ꢀNH2)
C H (ꢀNH2)
H (ꢀNH2)
A H (RH MBA)
B H (RH MBA)
C H (RH MBA)
1
Table 2. H Hyperfine Values Obtained by DFT Computa-
tions for the R,R-[Cu(1)] þ R-MBA and R,R-[Cu(1)] þ
S-MBA Adducts Corresponding to the Optimized Structures
A, B, and C (Figure 5)a
a The experimental 1H values obtained by ENDOR are also given. All A
values given in MHz. b Nucleus on the same side of the complex as the
MBA phenyl group; the hyperfine and quadrupole values were com-
puted with the basis set EPR-II. Error in A values for Himine, Ht-butyl, and
Hmethine = (0.02 MHz. Error in A values for HNH2 = (0.05 MHz.
DFT
experimental
nuclei
A1
A2
A3
A1
A2
A3
ENDOR experiments. To determine this interaction, additional
X-band HYSCORE experiments of [Cu(1)] in toluene and MBA
were undertaken (Supporting Information, Figure S14). Two
distinct peaks at ∼3.4 and 4.2) MHz were observed, centered
around the 13C Larmor frequency, and could thus stem from
nearby 13C nuclei in altered arrangement (as we proposed
previously).44 However, in principle, an appropriate combination
of 14N hyperfine and nuclear quadrupole couplings could also
lead to double-quantum cross-peaks at this position. In order to
test this, the HYSCORE spectrum of [Cu(1)] dissolved in
pyridine (Py) was also recorded (Figure S14, Supporting In-
formation). Analysis of the resulting spectra proved that the cross
peaks observed for [Cu(1)] þ MBA are indeed due to the weak
interaction with the amine nitrogen. The simulated parameters
are given in Table 3.
imine protons
[Cu(1)]
[Cu(2)]
18.5
18.4
17.2
16.6
16.6
17.2
17.7
16.6
17.2
16.1
22.1
22.45
21.0
21.3
21.3
[Cu(1)] þ MBA
[Cu(2)] þ MBA (HA)
(HB)
A H1 (imine)b
20.16 23.97 19.36
28.54 24.58 23.85
22.25 25.95 21.50
27.49 23.60 22.90
19.92 23.70 19.13
28.83 24.85 24.08
H2 (imine)
B H1 (imine)b
H2 (imine)
C H1 (imine)b
H2 (imine)
methine protons
[Cu(1)]
ꢀ1.41
5.54 ꢀ1.43
In summary, ENDOR and HYSCORE were employed to
examine the structure of the bound asymmetric adducts via
analysis and comparison of the spectra. Unfortunately, despite
A H1 (methine)b
H2 (methine)
B H1 (methine)b
H2 (methine)
C H1 (methine)b
H2 (methine)
ꢀ0.59
ꢀ0.05
0.03
7.26 ꢀ1.67
5.77 ꢀ0.67
7.50 ꢀ1.01
5.67 ꢀ1.51
6.91 ꢀ1.95
5.77 ꢀ0.71
1
the fact that the H ENDOR spectra of the imine protons did
reveal a lowering in the symmetry of the adducts, the hyperfine
data did not reveal any diastereomeric discrimination, as detected
by EPR. The data did, however, confirm the weak binding mode
of MBA to the Cu complexes (i.e., upon MBA binding the ligand
ꢀ0.58
ꢀ0.89
ꢀ0.00
1
14N couplings changed only slightly, the amine H and 14N
tert-butyl protons
[Cu(1)]
ꢀ1.64 ꢀ1.50
3.30
couplings were very small, and the ligand methine and tert-butyl
groups were hardly affected).
A H (tert-butyl)
ꢀ1. 11 ꢀ1.41
ꢀ1.08 ꢀ1.47
ꢀ1.07 ꢀ1.50
ꢀ1.15 ꢀ1.40
ꢀ1.06 ꢀ1.29
ꢀ1.11 ꢀ1.59
ꢀ1.21 ꢀ1.67
ꢀ1.06 ꢀ1.31
ꢀ1.10 ꢀ1.37
ꢀ1.08 ꢀ1.49
ꢀ1.09 ꢀ1.51
ꢀ1.10 ꢀ1.39
2.81
3.16
3.91
2.97
2.70
3.14
2.34
2.68
2.96
3.15
2.89
2.79
’ COMPUTATIONAL DFT ANALYSIS
The interaction of R- and S-MBA with R,R-[Cu(1)] was also studied
computationally. The metal complex was studied at the QM (BHandH)
level, while the aliphatic tert-butyl groups and cyclohexyl backbone were
B H (tert-butyl)
C H (tert-butyl)
studied at the MM level (UFF). The Cu
N
amine distance was initially
3 3 3
fixed at a relatively short 2.2 Å, and the bound MBA substrate was
allowed to rotate about the CuꢀN bond in 10ꢀ steps. The plot of the
relative energy versus the OꢀCu NꢀC dihedral angle for R,R-[Cu-
3 3 3
(1)] with R-MBA is shown in the Supporting Information (Figure S16).
Two clear energy minima were found for RR þ R corresponding to
6950
dx.doi.org/10.1021/ic200113u |Inorg. Chem. 2011, 50, 6944–6955

Inorganic Chemistry
ARTICLE
Table 3. Experimental (and Calculated) 14N Hyperfine and
Nuclear Quadrupole Values Obtained by ENDOR and
HYSCOREa
b
c
nuclei
[Cu(1)]I
[Cu(2)]I
A1
A2
A3
P1
P2
P3 e2qQ/hg ηh
50.5 37.4 38.5 ꢀ1.15 0.70 0.45 ꢀ2.3 0.2
50.4 36.6 38.0 ꢀ1.25 0.64 0.48 ꢀ2.5 0.15
[Cu(1)] þ MBA 49.7 35.9 36.4 ꢀ1.20 0.70 0.50 ꢀ2.4 0.2
[Cu(2)] þ MBA 50.0 34.2 38.0 ꢀ1.25 0.80 0.45 ꢀ2.5 0.0
N (MBA)d
A N1*
0.7e 0.8e 0.9f
3.3 0.9
34.37 48.05 35.79 1.01 ꢀ1.25 0.24
52.92 36.74 38.35 ꢀ1.30 1.02 0.28
34.90 49.83 36.60 1.00 ꢀ1.30 0.30
51.43 35.43 37.35 ꢀ1.35 1.02 0.33
33.62 47.12 35.06 1.01 ꢀ1.28 0.27
53.40 37.65 38.65 ꢀ1.29 1.02 0.27
ꢀ0.81 ꢀ0.85 0.29 1.29 0.93 ꢀ2.22
ꢀ0.75 ꢀ0.77 0.37 1.29 0.91 ꢀ2.20
ꢀ0.77 ꢀ0.78 0.39 1.29 0.89 ꢀ2.18
N2
B N1*
N2
C N1*
N2
A N (MBA)
B N (MBA)
C N (MBA)
a The corresponding 14N values obtained by DFT computations for the
R,R-[Cu(1)] þ R-MBA and R,R-[Cu(1)] þ S-MBA adducts are also
given, corresponding to the optimized structures labeled A, B, and C
(Figure 5). All A and P values given in MHz. b A values (0.2 MHz.
c P values (0.1 MHz. d Euler angles = [30 13 0]ꢀ ( 10ꢀ. e (0.1 MHz.
f (0.3 MHz. g (0.2 MHz. h (0.1 MHz. Only the values for N2 are
I
given. Those for N1 are identical; both A and P tensors are rotated in a
plane 90ꢀ around the A3 (P3) axis.
dihedral angles of j = ꢀ102ꢀ and 139ꢀ (definition of this angle is given
in the Supporting Information, Figure S15) with one minimum for RR þ S
possessing a dihedral angle of j = ꢀ21.9ꢀ. Free optimization of the energy
at these minima was then allowed, which resulted in slightly longer
Cu Namine distances of 2.255 and 2.253 Å for the RR þ R adducts
3 3 3
and 2.267 Å for the RR þ S adduct. These structures are hereafter labeled A,
B, and C respectively (Figure 5). Structures A and B were practically
isoenergetic (although there was a 2 kJ molꢀ1 difference between the two in
favor of A), while both were at least 7 kJmolꢀ1 lower in energy than all other
RR þ R orientations. Most importantly, structure C for the heterochiral RR
þ S was found to be 5 kJ molꢀ1 lower in energy compared to the
homochiral adducts RR þ R. Note that unrestricted optimization of the RR
þ S adduct led to a final structure C with the MBA phenyl orientation quite
similar to the one obtained for the most stable structure A of the RR þ R
adduct (Figure 5).
Although the inability of DFT to accurately quantify dispersion effects
renders it unsuitable for modeling πꢀπ interactions, the computational
results indicate that the adduct could be likely stabilized by an offset
πꢀπ interaction between the phenyl ring of MBA and the imine group
of the complex. This is in keeping with the πꢀπ rules outlined by
Hunter and Sanders22a and could explain the lower symmetry of the
adducts observed by EPR and ENDOR. Recent work on benzene,
pyridine, and DNA base dimers demonstrated minima on the potential
energy surfaces with the BHandH functional.34 They attributed this
ability to qualitatively reproduce geometric details to a fortuitous
cancellation of errors within the functional and demonstrated that it
applied to a range of πꢀπ interacting systems, reproducing energies
generally to within (2 kJ molꢀ1 of the MP2 values. For these reasons
this functional was used here with the BHandH to model the πꢀπ
interactions. However, no improvements were observed in the calcula-
tions of the structures A, B, and C, so we did not consider further the
stabilization role of πꢀπ interactions in our interpretations.
Figure 5. QMMM-optimized structures for R-MBA with R,R-[Cu(1)]
from the constrained scan reported in Figure S14, Supporting Informa-
tion, giving structure A (Cu N = 2.255 Å) and structure B (Cu
3 3 3
3 3 3
N = 2.253 Å). Structure C is optimized from the equivalent scan for
S-MBA with R,R-[Cu(1)], giving Cu N = 2.267 Å. Atom colors: for
3 3 3
clarity, C ([Cu(1)]) and C(MBA) are colored gray and green, respec-
tively; N, blue; O, red; H, white; Cu, pink.
parameters are presented in Tables 1ꢀ3. Despite the fact that current
state-of-the-art DFT methods still struggle to exactly reproduce the g
and metal hyperfine values of metal complexes,45 the agreement
between experimental g/ACu values and the calculated parameters
(Table 1) is satisfactory. The in-plane principal g and CuA tensor axes
In a next step, the spin Hamiltonian parameters were computed for
the three optimized DFT structures A, B, and C. The relevant EPR
6951
dx.doi.org/10.1021/ic200113u |Inorg. Chem. 2011, 50, 6944–6955

Inorganic Chemistry
ARTICLE
were found to be colinear, pointing approximately along the OꢀCuꢀN axes.
The calculated ligand 14N hyperfine and quadrupole parameters are
listed in Table 3. The ligand (1) nitrogens, labeled N1 and N2, were
found to be inequivalent in terms of their hyperfine and nuclear
quadrupole parameters. N1 refers to the ligand 14N nucleus under the
phenyl ring of the bound MBA substrate (i.e., on the same side of the
[Cu(1)] complex; see Figure 5), while N2 refers to the 14N nucleus on
the opposite side of the complex from the MBA. The N1 and N2
hyperfine and nuclear quadrupole tensors are approximately colinear
with the g tensor axes, with the largest hyperfine value lying along the
CuꢀN axes. The agreement between experimental data and theory
(Table 3) is very good. Interestingly, while slight inequivalencies
between the two ligand 14N nuclei revealed via DFT, these difference
were not directly observed experimentally (although a slight broadening
of the peaks was indeed detected in the 14N ENDOR spectra, suggestive
of in-equivalent 14N nuclei; Supporting Information). The A and P
parameters for the 14N nucleus of the weakly interacting MBA substrate
were detected via X-band HYSCORE, and the results are also in rea-
sonable agreement with the computed values (Table 3).
In the present system, differences in diastereomeric adducts
were sufficiently pronounced that enantiomer discrimination was
directly manifested in the W-band EPR spectra. The heterochiral
CuꢀMBA combinations (RR-S and SS-R) were found to be the
preferred adducts with selectivities of 86 ( 5% and 57 ( 5% for
[Cu(1)] and [Cu(2)], respectively. While the role of H bonds
was shown to be instrumental in controlling the stereoselectiv-
ities of the [VO(1)]ꢀepoxides system,20 the mode of stereo-
chemical communication between [Cu(1,2)] and R-/S-MBA
must also be rationalized.
According to the computational DFT results for the homo-
chiral (R,R-[Cu(1)] þ R-MBA) and heterochiral (R,R-[Cu(1)] þ
S-MBA) adducts, the bulky MBA substrate was found to bind to
the bulky [Cu(1)] complex. The heterochiral adduct was en-
ergetically favored (by only 5 kJ molꢀ1) in agreement with the
experimental observation. Two discrete energy minima were
identified in the homochiral adducts (structures A and B;
Figure 5) in which the MBA-phenyl group was positioned over
either phenyl ring of ligand (1). In the heterochiral adduct
(structure C; Figure 5) a slightly different alignment of the
MBA-phenyl ring was observed. Nevertheless, in all cases, the
adducts were found to be less energetically preferred when
the MBA-phenyl ring was positioned over the tert-butyl groups
in [Cu(1)] due to steric hindrance. Furthermore, the methyl
group attached to the asymmetric carbon atom of MBA was
always found to point ‘upward’ and away from the ligand plane in
all adducts (Figure 5). As a result, the single R-proton attached to
this same asymmetric carbon atom of MBA points ‘downward’
toward the ligand. This has important consequences for control-
ling the stereoselective binding in homo- vs heteroadducts as
discussed below.
The slight lowering in the symmetry of the adduct upon MBA
coordination was detected experimentally via the inequivalency of the
imine
resulting A values for HA,B
(Figure 4 and Table 2). This was also
confirmed in the computed A values for both the imine and the methine
imine
protons. For example, in structure A, two sets of A values for HA,B
were calculated (Table 2). The computed A values for Hmethine were also
found to be inequivalent; this was not be detected experimentally and is
possibly due to a slight distribution in the conformations. Nevertheless,
the agreement between the experimental and calculated A values for
H
imine and Hmethine are very good. A range of A values was extracted from
the computational studies for the inner tert-butyl groups, since the
absolute values obtained depended on the adopted rotational conforma-
tion. The hyperfine values for the four closest Ht-butyl protons are listed
in Table 2, while all the remaining protons produced smaller A values.
The experimental values represent an excellent average of the four
closest Ht-butyl hyperfine values listed in Table 2.
In structure A, the closest contact points of MBA with [Cu(1)]
(apart from the Cu
N
amine contact) was between the methine
3 3 3
ligand proton and both the R-proton and o-phenyl proton of
MBA (Hmethine HR or ortho distances of 2.40 and 2.58 Å res-
According to the calculations, two distinct sets of hyperfine values
were found for protons originating from the bound MBA substrate;
these included the amine ꢀNH2 protons and the R-proton attached to
the asymmetric carbon atom of MBA (Scheme 1). For example, the
calculated A values for the ꢀNH2 protons for structure A are somewhat
different from the experimental values (Table 2). This is not surprising
since these hyperfine values will be very sensitive to small changes in
the MBA ligand orientation and the CuꢀNMBA distance. Again, the
calculated asymmetry of the NH2 protons was not observed in the
experiment, in part due to a possible distribution of conformers and also
due to the broader line widths for these protons in the ENDOR spectra.
A smaller R-proton hyperfine was also calculated (Table 2), although
this hyperfine was not observed in the experimental ENDOR spectra, as
the peaks would be superimposed on other ligand-derived peaks, making
any useful assignment very difficult.
3 3 3
pectively). In this homochiral adduct, the arrangement is such
that the R-proton of R-MBA points toward the methine proton in
R,R-[Cu(1)] (Figure 5A). However, in the heterochiral adduct,
this R-proton of S-MBA points away from the methine proton in
R,R-[Cu(1)] (Figure 5C). As a result, the closest point of contact
with S-MBA is now only between the methine ligand proton and
the o-phenyl proton of MBA (distance of 2.47 Å). This slight
difference in steric crowding caused by the R-proton of MBA
appears to be sufficient to tip the balance of selectivity in favor of
the heterochiral adduct.
It should also be mentioned that although there was little
difference in energy between structures A and B (for the homo-
chiral adduct), close inspection of the structures reveals that the
MBA-phenyl ring is slightly displaced over the cyclohexyl ring of
(1) in structure B. As a result, more of the MBA-phenyl ring
protons now overlap the staggered cyclohexyl ring protons, and
this creates a small but unfavorable amount of steric crowding.
Such an outcome is not observed in structure A, where only two
points of closest contact are observed (see above). This may be
responsible for the slight preference for the amine to bind on one
side of the complex in R,R-[Cu(1)] þ R (structure A).
’ DISCUSSION
Stereoselective Binding of MBA with [Cu(1,2)]. In previous
studies, we investigated the discrimination of chiral epoxides20 by
chiral [VO(1)] complexes. The [VO(1)] þ epoxide adducts
were very weak, but differences in the EPR and ENDOR spectra
of the homochiral and heterochiral adducts were readily iden-
tifiable.20 The origin of this enantioselectivity was based upon a
series of weak H bonds, augmented by an electrostatic interac-
tion, between the epoxide substrate and the [VO(1)] ligand.
Crucially these hydrogen bonds facilitated the overall orientation
of the epoxide between the metal center and the chiral salen
backbone.20
It is important to emphasize that the MBA substrate is
effectively trapped in one of two positions when bound to
[Cu(1)], for example, on one side of the chiral salen complex
as shown in Figure 5 (A or B). This is due to the inability of the
MBA-phenyl ring to easily pass over the cyclohexyl ring due to
the axial methine proton, which is virtually perpendicular to the
6952
dx.doi.org/10.1021/ic200113u |Inorg. Chem. 2011, 50, 6944–6955

Inorganic Chemistry
ARTICLE
salen ligand plane. This methine proton, on the asymmetric
carbon center of the diamine backbone, thus creates a barrier to
MBA rotation. While this ligand proton is not directly involved in
the stereoselectivity for MBA binding (as occurs for chiral
discrimination via H bonds in [VO(1)] þ epoxides),20 it does
indirectly influence the mode of binding by restricting easy
movement of the substrate from structure A to B. Furthermore,
due to steric interactions with the R-proton attached to the asym-
metric carbon atom of R-MBA, this methine proton is also
indirectly responsible for favoring the heterochiral structure C
over the homochiral structure A.
Finally, removal of the tert-butyl groups in [Cu(2)] enables the
chiral MBA substrates to adopt more binding orientations
relative to [Cu(1)]. In particular, the MBA substrate does not
experience any steric repulsion caused by the bulky ligand groups
in [Cu(1)]; as more orientations are accessible, the stereoselec-
tivity in the binding obviously decreases substantially to only
57 ( 5%. Despite the fact that the MBA-phenyl ring can poten-
tially overlap and π stack with the ligand phenyl rings (Figure 5)
and thus facilitate further stabilization of the adducts, DFT was
unable to accurately quantify these possible contributions.
Implications of Findings for Asymmetric Homogeneous
Catalysis. One of the unique features of privileged chiral catalysts
is their ability to deliver a broad applicability for so many diverse
reactions. Clearly certain key structural features of the catalysts
are important to achieve these high levels of enantioselectivities.
Bulky framework substituent’s are known to prevent stabilization
of transition states, particularly in asymmetric catalysis, since the
transition states for the two diastereomers have similar energies
(although one will have a slightly lower kinetic barrier compared
to the other). This is particularly true in chiral metal salen
complexes, whereby the enantioselectivities of the asymmetric
epoxidation reactions using the Mn(salen) catalyst depends on
the bulky substituents at the 3,30 and 5,50 positions.9c,46 These
bulky substituents not only prevent stabilization of transition
states but also play another important role by blocking and
regulating the orientation of the incoming substrates, creating a
high diastereofacial preference.9c,46
It is clear from the results presented here that strong diaster-
eomeric discrimination of one specific enantiomer of a substrate
occurs in the R,R-[Cu(1)] complex (86 ( 5% for RR-S and
SS-R). In the absence of these bulky substituents (as in [Cu(2)])
the diastereomeric discrimination of the chiral amine substrates
is substantially reduced (57 ( 5% for RR-S and SS-R). This
indicates that the tert-butyl groups are directly implicated in
regulating chiral recognition and can help in the differentiation
between different stereoisomers of the amine. As the DFT results
have shown, subtle steric hindrance between substrate substitu-
ents (such as MBA R-proton) and ligand substituents (such as
tert-butyl or Hmethine) are responsible for destabilizing one
diastereomeric adduct in favor of another. While the presence
of such diastereomeric adducts are often presumed as mechan-
istic intermediates, they are rarely observed directly in cases of
weak complexꢀsubstrate interactions. The results reported here
therefore demonstrate the useful role of W-band EPR in probing
such diastereomeric adducts which may be of direct relevance to
studies in homogeneous asymmetric catalysis.
revealed by CW EPR and DFT. The spectroscopic data revealed
that only one MBA substrate bound weakly to the copper com-
plexes. Diastereomeric discrimination of MBA enantiomers was
directly observed by W-band EPR, revealing an 86:14 preference
for the heterochiral adducts (RR-S and SS-R) in [Cu(1)]; this
diminished to 57:43 in favor of the heterochiral adducts in
[Cu(2)]. The symmetry of the [Cu(1,2)] þ MBA adducts was
lowered compared to the uncoordinated [Cu(1,2)] complexes;
as reflected in the rhombic g distortions, this lower symmetry
resulted from the preferential binding and orientation of the
MBA substrate onto one side of the chiral Cu complex.
While the W-band EPR data revealed the selectivity of MBA
binding, the origin of this selectivity was explained by DFT. The
results revealed that the bulky phenyl ring of MBA destabilized
formation of any adduct which placed the MBA-phenyl ring over
the tert-butyl groups at positions 3,30 and 5,50 of the complex.
Instead, steric hindrance between the complex and the substrate
was minimized when the MBA-phenyl ring was positioned
over the phenyl rings of the [Cu(1)] complex. Two stabilization
sites were identified in the homochiral adduct R,R-[Cu(1)] þ
R-MBA (labeled A and B); structure A was slightly preferred
by 2 kJ molꢀ1 due to the small unfavorable steric interactions
between the MBA-phenyl ring and the ligand cyclohexyl ring
that occurred in B. However, the most stable site was found
for the heterochiral adduct R,R-[Cu(1)] þ S-MBA (labeled C),
in agreement with the experiments. This site C was slightly
preferred by 5 kJ molꢀ1 compared to the homochiral adduct sites
A and B. In this heterochiral case, the R-proton of S-MBA was
found to point away from the ligand methine proton; the reverse
situation occurred in the homochiral adducts. These results
reveal how very weak, and indeed subtle, outer-sphere interac-
tions are responsible for controlling stereoselective binding in
chiral complexꢀchiral substrate interactions.
’ ASSOCIATED CONTENT
S
Supporting Information. Further experimental details,
b
X-band EPR spectra of [Cu(2)] in MBA, determination of
1
the coordination number of MBA molecules, supporting H
and 14N CW and pulsed ENDOR spectra, HYSCORE, and
additional DFT details. This material is available free of charge
via the Internet at http://pubs.acs.org.
’ AUTHOR INFORMATION
Corresponding Author
*E-mail: MurphyDM@cardiff.ac.uk; sabine.vandoorslaer@
ua.ac.be.
’ ACKNOWLEDGMENT
This work was supported by EPSRC funding (EP/E030122),
by the Hercules Foundation, Flanders (contract AUHA013), and
a NOI-BOF grant of the University of Antwerp (to S.V.D.).
’ REFERENCES
(1) (a) Heitbaum, M.; Glorius, F.; Escher, I. Angew. Chem., Int. Ed.
2006,45, 4732–4762. (b) Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.;
In Comprehensive Asymmetric Catalysis; Springer: New York, 1999; Vols.
1ꢀ3. (c) Ojima, I., Ed.; In Catalytic Asymmetric Synthesis, 2nd ed.;
Wiley-VCH: New York, 2000.
’ CONCLUSIONS
The asymmetric interaction of the chiral amines (R-/S-
methylbenzylamine) with the chiral [Cu(1,2)] complexes was
6953
dx.doi.org/10.1021/ic200113u |Inorg. Chem. 2011, 50, 6944–6955

Inorganic Chemistry
ARTICLE
(2) (a) Noyori, R. Angew. Chem., Int. Ed. 2002, 41, 2008–2022.
(b) Knowles, W. S. Angew. Chem., Int. Ed. 2002, 41, 1998–2007.
(c) Sharples, K. B. Angew. Chem., Int. Ed. 2002, 41, 2024–2032.
(3) (a) Walsh, P. J.; Kozlowski, M. C. Fundamentals of asymmetric
catalysis; University Science Books: Sausalito, CA, 2009. (b) In Asymmetric
synthesisꢀThe essentials, 2nd ed.; Christmann, M., Brase, S., Eds.; Wiley-VCH:
New York, 2008.
(4) Yoon, T. P.; Jacobsen, E. M. Science 2003, 299, 1691–1693.
(5) (a) Jacobsen, E. N. In Catalytic Asymmetric Synthesis; Ojima, I.,
Ed.; VCH: Weinheim, 1993. (b) Jacobsen, E. N. In Comprehensive
Organometallic Chemistry II; Wilkinson, G.; Stone, F.G.A.; Abel, E. W.,
Hegedus, L. S., Eds.; Pergamon Press, New York, 1995; Vol. 12.
(c) Katsuki, T. Coor. Chem. Rev. 1995, 140, 189–214. (d) Dalton,
C. T.; Ryan, K. M.; Wall, V. M.; Bousquet, C.; Gilheany, D. G. Top. Catal.
1998, 5, 75. (e) Canalai, L.; Sherrington, D. C. Chem. Soc. Rev. 1999,
28, 85–93.
(6) (a) Costas, M.; Mehn, M. P.; Jensen, M. P.; Que, L., Jr. Chem.
Rev. 2004, 104, 939–986. (b) Di Mauro, E. F.; Mamai, A.; Kozlowski,
M. C. Organometallics 2003, 22, 850–855.
(7) (a) Collman, J. P.; Zhang, A.; Lee, V. J.; Uffelman, E. S.;
Brauman, J. I. Science 1993, 261, 1404–1411. (b) Sono, M.; Roach,
M. P.; Coulter, E. D.; Dawson, J. H. Chem. Rev. 1996, 96, 2841–2888.
(8) Venkataramanan, N. S.; Kuppuraj, G.; Rajagopal, S. Coord. Chem.
Rev. 2005, 249, 1249–1268.
(21) (a) Fallis, I. A.; Murphy, D. M.; Willock, D. J.; Tucker, R. J.;
Farley, R. D.; Jenkins, R.; Strevens, R. J. Am. Chem. Soc. 2004,
126, 15660–15661. (b) Murphy, D. M.; Fallis, I. A.; Landon, J.; Willock,
D. J.; Carter, E.; Van Doorslaer, S.; Vinck, E. Phys. Chem. Chem. Phys.
2009, 11, 6757–6769. (c) Murphy, D. M.; Fallis, I. A.; Willock, D. J.;
Landon, J.; Carter, E.; Vincks, E. Angew. Chem., Int. Ed. 2008,
47, 1414–1416. (d) Carter, E.; Murphy, D. M.; Fallis, I. A.; Willock,
D. J.; Van Doorslaer, S.; Vinck, E. Chem. Phys. Lett. 2010, 486, 74–79.
(22) (a) Hunter, C. A.; Sander, J. K. M. J. Am. Chem. Soc. 1990,
112, 5525–5534. (b) Miyamura, K.; Mihara, A.; Fujii, T.; Gohshi, Y.;
Ishii, Y. J. Am. Chem. Soc. 1995, 117, 2371–2318. (c) Askew, B.; Ballester,
P.; Buhr, C.; Jeong, K. S.; Jones, S.; Parris, K.; Williams, K.; Rebek, J.
J. Am. Chem. Soc. 1989, 111, 1082–1090. (d) Abraham, R. J.; Eivazi, F.;
Pearson, H.; Smith, K. M. J. Chem. Soc., Chem. Commum. 1976, 699–701.
(e) Clever, G. H.; Carell, T. Angew. Chem., Int. Ed. 2007, 46, 250–253.
(23) (a) Lyubenova, S.; Maly, T.; Zwicker, K.; Brandt, U.; Ludwig,
B.; Prisner, T. Acc. Chem. Res. 2010, 43, 181–189. (b) Goldfarb, D.;
Arieli, D. Ann. Rev. Biophys. Biomol. Struct. 2004, 33, 441–468.
(c) Astashkin, A. V.; Raitsimring, A. M.; Walker, F. A.; Rensing, C.;
McEvoy, M. M. J. Biol. Inorg. Chem. 2005, 10, 221–230. (d) Derose, V. J.;
Hoffman, B. M. Methods Enzymol. 1995, 246, 554–589. (e) Ames,
W. M.; Larsen, S. C. J. Biol. Inorg. Chem. 2009, 14, 547–557.
(24) (a) Yoon, J.; Solomon, E. I. Coor. Chem. Rev. 2007,
251, 379–400. (b) Finazzo, C.; Calle, C.; Stoll, S.; Van Doorslaer, S.;
Schweiger, A. Phys. Chem. Chem. Phys. 2006, 8, 1942–1953. (c) Wagner,
M. R.; Walker, F. A. Inorg. Chem. 1983, 22, 3021–3028. (d) Shao, J. L.;
Steene, E.; Hoffman, B. M.; Ghosh, A. Eur. J. Inorg. Chem. 2005,
8, 1609–1615. (e) Zhao, M.; Stern, C.; Barrett, A. G. M.; Hoffman,
B. M. Angew. Chem. 2003, 42, 462–465.
(9) (a) Katsuki, T. Adv. Synth. Catal. 2002, 344, 131–147.
(b) Katsuki, T. Curr. Chem. Org. 2001, 63. (c) Palucki, M.; Finney,
N. S.; Pospisil, P. J.; Guler, M. L.; Ishida, T.; Jacobsen, E. N. J. Am. Chem.
Soc. 1998, 120, 948–954. (d) Tokunaga, M.; Larrow, J. F.; Kakiuchi, F.;
Jacobsen, E. N. Science 1997, 277, 936–938.
(10) Jacobsen, E. N. Acc. Chem. Res. 2000, 33, 421–431.
(11) (a) Fraile, J. M.; Garcia, J. I.; Jimenez-Oses, G.; Mayoral, J. A.;
Roldan, M. Organometallics 2008, 27, 2246–2251. (b) Di Tommaso, D.;
French, S. A.; Zanotti-Gerosa, A.; Hancock, F.; Palin, E. J.; Catlow,
C. R. A. Inorg. Chem. 2008, 47, 2674–2687. (c) Ihori, Y.; Yamashita, Y.;
Ishitani, H.; Kobayashi, S. J. Am. Chem. Soc. 2005, 44, 15528–15535.
(d) Brandt, P.; S€odergren, M. J.; Gandersson, P.; Norrby, P.-O. J. Am.
Chem. Soc. 2000, 122, 8013–8020.
(25) (a) Bunce, S.; Cross, R. J.; Farrugia, L. J.; Kunchandy, S.;
Meason, L. L.; Muir, K. W.; Donnell, M. O; Peacock, R. D.; Stirling, D.;
Teat, S. J. Polyhedron 1998, 17, 4179–4187. (b) Bernardo, K.; Leppard,
S.; Robert, A.; Commenges, G.; Dahan, F.; Meunier, B. Inorg. Chem.
1996, 35, 387–396.
(26) Stoll, S.; Schweiger, A. J. Magn. Reson. 2006, 178, 42–55.
(27) H€ofer, P.; Grupp, A.; Nebenf€uhr, H.; Mehring, M. Chem. Phys.
Lett. 1986, 132, 279–282.
(12) (a) Van Doorslaer, S.; Caretti Giangaspro, I.; Fallis, I. A.;
Murphy, D. M. Coor. Chem. Rev. 2009, 253, 2116–2130. (b) Carter,
E.; Murphy, D. M. In Spectroscopic Properties of Inorganic and Organo-
metallic Compounds; Douthwaite, R., Duckett, S., Eds.; RSC: Cambridge,
2009; Vol. 4, pp 355ꢀ384.
(28) Madi, Z. L.; Van Doorslaer, S.; Schweiger, A. J. Magn. Reson.
2002, 154, 181–191.
(29) Davies, E. R. Phys. Lett. A 1974, A 47, 1–2.
(30) (a) Schosseler, P.; Wacker, T.; Schweiger, A. Chem. Phys.
Lett. 1994, 224, 319–324. (b) Schweiger, A.; Jeschke, G. Principles
of Pulse Electron Paramagnetic Resonance; Oxford University Press:
Oxford, 2001.
(31) Vreven, T.; Morokuma, K. J. Comput. Chem. 2000, 21, 1419–
1432.
(13) (a) Bolm, C.; Martin, M.; Gescheidt, G.; Palivan, C.; Stanoeva,
T.; Bertagnolli, H.; Feth, M.; Schweiger, A.; Mitrikas, G.; Harmer, J.
Chem.—Eur. J. 2007, 13, 1842–1850. (b) Bolm, C.; Martin, M.;
Gescheidt, G.; Palivan, C.; Neshchadin, D.; Bertagnolli, H.; Feth, M.;
Schweiger, A.; Mitrikas, G.; Harmer, J. J. Am. Chem. Soc. 2003, 125,
6222–6227.
(32) Frisch, M. J.; et al. Gaussian 03, Revision C.02; Gaussian, Inc.:
Wallingford, CT, 2004.
(14) Srinivasan, K.; Michaud, P.; Kochi, K. J. J. Am. Chem. Soc. 1986,
108, 2309–2320.
(33) Rappꢀe, A. K.; Casewit, C. J.; Colwell, K. S.; Goddard, W. A., III;
Skiff, W. M. J. Am. Chem. Soc. 1992, 114, 10024–10035.
(15) (a) Bryliakov, K. P.; Babushkin, D. E.; Talsi, E. P. J. Mol. Catal. A
2000, 158, 19–35. (b) Bryliakov, K. P.; Khavrutskii, I. V.; Talsi, E. P.;
Kholdeeva, O. A. React. Kinet. Catal. Lett. 2000, 71, 183–191.
(c) Bryliakov, K. P.; Kholdeeva, O. A.; Vanina, M. P.; Talsi, E. P.
J. Mol. Catal. A 2002, 178, 47–53.
(34) Waller, M. P.; Robertazzi, A.; Platts, J. A.; Hibbs, D. E.;
Williams, P. A. J. Comput. Chem. 2006, 27, 491–504.
(35) Lee, C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789.
(36) (a) Neese, F. J. Chem. Phys. 2001, 115, 11080–11096.
(b) Neese, F. J. Phys. Chem. A 2001, 105, 4290–4299. (c) Neese, F.
J. Chem. Phys. 2003, 118, 3939–3948. (d) Neese, F. J. Chem. Phys. 2005,
122, 34107(1)–34107(13).
(16) Campbell, K. A.; Lashley, M. R.; Wyatt, J. K.; Nantz, M. H.;
Britt, R. D. J. Am. Chem. Soc. 2001, 123, 5710–5719.
(17) Hsieh, W.-Y.; Pecoraro, V. L. Inorg. Chim. Acta 2002, 341,
113–117.
(18) (a) Bryliakov, K. P.; Lobanova, M. V.; Talsi, E. P. J. Chem. Soc.,
Dalton Trans. 2002, 2263–2265. (b) Bryliakov, K. P.; Talsi, E. P. Inorg.
Chem. 2003, 42, 7258–7265.
(19) (a) Vinck, E.; Murphy, D. M.; Fallis, I. A.; Strevens, R. R.; Van
Doorslaer, S. Inorg. Chem. 2010, 49, 2083–2092. (b) Vinck, E.; Van
Doorslaer, S.; Murphy, D. M.; Fallis, I. A. Chem. Phys. Lett. 2008,
464, 31–37. (c) Vinck, E.; Murphy, D. M.; Fallis, I. A.; Van Doorslaer, S.
Appl. Magn, Reson. 2010, 37, 289–303.
(37) This basis is based on the TurboMole DZ basis developed
by Ahlrichs and co-workers and obtained from the basis set library
under ftp.chemie.unikarlsruhe.de/pub/basen. Unpublished results by
R. Ahlrichs and co-workers.
(38) Barone, V. In Recent Advances in Density Functional Methods;
Chong, D. P., Ed.; World Scientific Publ. Co.: Singapore, 1996; p 287.
(39) (a) Schaefer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992,
97, 2571–2577. (b) The Ahlrichs (2df,2pd) polarization functions were
obtained from the TurboMole basis set library under ftp.chemie.
uni-karlsruhe.de/pub/basen. Unpublished results by R. Ahlrichs and
co-workers.
(20) Movassaghi, M.; Jacobsen, E. M. Science 2002, 298, 1904–1905.
6954
dx.doi.org/10.1021/ic200113u |Inorg. Chem. 2011, 50, 6944–6955

Inorganic Chemistry
ARTICLE
(40) Sinnecker, S.; Rajendran, A.; Klamt, A.; Diedenhof, M.; Neese,
F. J. Phys. Chem. A 2006, 110, 2235–2245.
(41) (a) Schweiger, A. Struct. Bonding (Berlin) 1982, 51, 1. (b) Hs.,
H.; G€unthard, A.; Schweiger Chem. Phys. 1978, 32, 35–61. (c) Pilbow,
J. R. Transition Ion Electron Paramagnetic Resonance; Oxford Science
Publications: Oxford, 1990.
(42) Kita, S.; Hashimoto, M.; Iwaizumi, M. Inorg. Chem. 1979, 18,
3432–3438.
(43) (a) Pratt, R. C.; Stack, T. D. P. J. Am. Chem. Soc. 2003,
12, 8716–8717. (b) Pratt, R. C.; Mirica, L. M.; Stack, T. D. P. Inorg.
Chem. 2004, 43, 8030–8039. (c) Pratt, R. C.; Stack, T. D. P. Inorg. Chem.
2005, 44, 2367–2375.
(44) Van Doorslaer, S.; Murphy, D. M.; Fallis, I. A. Res. Chem.
Intermed. 2007, 33, 807–823.
(45) (a) Waller, M. P.; Robertazzi, A.; Platts, J. A.; Hibbs, D. E.;
Williams, P. A. J. Comput. Chem. 2006, 27, 491–504. (b) Neese, F. J. Biol.
Inorg. Chem. 2006, 11, 702–711. (c) Neese, F.; Munzarova, M. L. In
Calculation of NMR and EPR parameters; Kaupp, M., Buhl, M., Malkin,
V. G., Eds.; Wiley-VCH: Weinheim, 2004; p 21. (d) Stein, M.; van
Lenthe, E.; Baerends, E. J.; Lubitz, W. J. Am. Chem. Soc. 2001,
123, 5839–5840. (e) Neese, F. Coord. Chem. Rev. 2009, 256, 526–563.
(46) (a) Zhang, W.; Loebach, J. L.; Wilson, S. R.; Jacobsen, E. N.
J. Am. Chem. Soc. 1990, 112, 2801–2803. (b) McGarrigle, E. M.;
Gilheany, D. G. Chem. Rev. 2005, 105, 1563–1602.
6955
dx.doi.org/10.1021/ic200113u |Inorg. Chem. 2011, 50, 6944–6955