Open Access Article. Published on 04 December 2017. Downloaded on 8/6/2019 6:35:12 PM.
This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.
Chemical
Science
View Article Online
EDGE ARTICLE
Cite this: Chem. Sci., 2018, 9, 1136
View Journal | View Issue
Enantiomeric resolution and X-ray optical activity
of a tricobalt extended metal atom chain†
Anandi Srinivasan,‡ab Miguel Cortijo, ‡abcd Vladimir Bulicanu,ab Ahmad Naim,
Rodolphe Clérac, ab Philippe Sainctavit,*e Andrei Rogalev,f Fabrice Wilhelm,f
Patrick Rosa cd and Elizabeth A. Hillard *ab
cd
A simple procedure based on anion exchange was employed for the enantiomeric resolution of the
extended metal atom chain (EMAC) [Co3(dpa)4(MeCN)2]2+. Use of the chiral salt (NBu4)2[As2(tartrate)2],
(L-1
or
D-1),
resulted
in
the
selective
crystallization
of
the
EMAC
enantiomers
as
[D-
Co3(dpa)4(MeCN)2](NBu4)2[L-As2(tartarte)2]2, (D-2) and [L-Co3(dpa)4(MeCN)2](NBu4)2[D-As2(tartrate)2]2
(L-2), respectively, in the P4212 space group, whereas a racemic mixture of 1 yielded [Co3(dpa)4(MeCN)2]
[As2(tartrate)2]$2MeCN (rac-3), which crystallized in the C2/c space group. The local electronic and
magnetic structure of the EMAC enantiomers was studied, exploiting a variety of dichroisms in single
crystals. A strong linear dichroism at the Co K-edge was observed in the orthoaxial configuration,
whereas it vanished in the axial orientation, thus spectroscopically confirming the D4 crystal symmetry.
Received 21st September 2017
Accepted 4th December 2017
DOI: 10.1039/c7sc04131d
rsc.li/chemical-science
Compounds D-2 and L-2 are shown to be enantiopure materials as evidenced by mirror-image natural
circular dichroism spectra in the UV/vis in solution and in the X-ray range at the Co K-edge in single
crystals. The surprising absence of detectable X-ray magnetic circular dichroism or X-ray magnetochiral
dichroism signals at the Co K-edge, even at low temperature (3 K) and a high magnetic field (17 T), is
ascribed to a strongly delocalized spin density on the tricobalt core.
Introduction
Extended Metal Atom Chains (EMACs) are a fascinating class of
linear polynuclear coordination complexes, which have served
as a rich playing eld for the study of metal–metal bonding,
magnetism and electrical conductivity.1 In addition, most
EMACs demonstrate helicoidal chirality arising from steric
hindrance between the protons in the 3-position of the pyridine
groups (Chart 1). These complexes almost exclusively crystallize as racemates, and chiral resolution has been accomplished in only a few isolated cases. For example,
spontaneous resolution afforded crystals of homochiral
[Co3(dpa)4(MeCN)2](PF6)2$MeCN$2Et2O (where dpa ¼ the anion
of 2,20 -dipyridylamine), but only as a minor product.2 These
crystals were then painstakingly separated using X-ray diffraction to determine the handedness of each crystal, in order to
then obtain circular dichroism (CD) measurements on an
enantiopure sample. In another study, homochiral samples of
a mixture of [Ni3(dpa)4Cl2] and [Ni3(dpa)4ClOH] were obtained
by chromatographic separation on a chiral stationary phase,
while the cobalt and chromium analogues were unstable
to chromatography.3 More recently, naphthyridine ligands
CNRS, CRPP, UPR 8641, F-33600 Pessac, France. E-mail: hillard@crpp-bordeaux.
cnrs.fr
a
b
c
Univ. Bordeaux, CRPP, UPR 8641, F-33600 Pessac, France
CNRS, ICMCB, UPR 9048, F-33600 Pessac, France
Univ. Bordeaux, ICMCB, UPR 9048, F-33600 Pessac, France
d
Institut de Mineralogie, de Physique des Materiaux et de Cosmochimie, UMR 7590,
CNRS, UPMC, IRD, MNHN, F-75005 Paris, France. E-mail: Philippe.Sainctavit@
impmc.upmc.fr
e
European Synchrotron Radiation Facility (ESRF), F-38043 Grenoble, France
f
† Electronic supplementary information (ESI) available: Bond distances and
angles for 2 and 3, thermal ellipsoid plots, packing diagrams, PXRD data,
magnetization curves, thermogravimetric analysis, and additional XNCD and
XMCD plots and CD spectra. CCDC 1574514–1574516 and 1588703. For ESI and
crystallographic data in CIF or other electronic format see DOI:
10.1039/c7sc04131d
‡ Both authors contributed equally to this work.
1136 | Chem. Sci., 2018, 9, 1136–1143
Chart 1 (Left) Crystallographic model of [Co3(dpa)4Cl2]; cobalt (dark
blue), nitrogen (pale blue), chloride (green), carbon (gray), hydrogen
(pink). (Right) Helicoidal wrapping of the dpa ligands; arrow indicates
H-atom repulsions.
This journal is © The Royal Society of Chemistry 2018
View Article Online
Open Access Article. Published on 04 December 2017. Downloaded on 8/6/2019 6:35:12 PM.
This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.
Edge Article
incorporating a chiral camphorsulfonyl group were conceived
for the enantiopure synthesis of Ni5 chains.4
The helicoidal chirality found in EMACs has several interesting features. UV/vis circular dichroism and polarimetry
studies showed that the above-mentioned enantiomers of tricobalt, trinickel and pentanickel EMACs do not interconvert in
solution, which is rather unusual for labile rst-row transition
metal complexes. This can be ascribed to a high energy barrier
to racemization engendered by the polynuclear helix. Moreover,
the optical activity for an acetonitrile solution of the mixture of
nickel EMACs obtained by chromatography was found to be
remarkably high, with a specic rotation of [a]22
D z 5000
deg mL g 1 dm 1.3b This value is on the same order of magnitude as that of the helicenes, the current record-holders for high
specic rotation.5
Natural circular dichroism, rst observed in 1895 by Aimé
Cotton, is one of the most powerful tools for obtaining stereochemical information. Circular dichroism measurements have
been only recently extended to X-ray absorption spectroscopy,
which is inherently element selective. It was rst achieved for Xray magnetic circular dichroism (XMCD), which has become
a workhorse technique for the exploration of the magnetic
properties of absorbing atoms in magnetic materials. The rst
observation of X-ray Natural Circular Dichroism (XNCD) was
made in 1998 on single crystals of a-LiIO3 and Na3[Nd(digly)3]$
2NaBF4$6H2O.6 Since then, a few other molecular systems have
been studied, such as [Co(en)3Cl3]$NaCl$6H2O and the 1D
and
coordination
polymers
[Co(hfac)2NITPhOMe]N
[Ln(Hnic)(nic)2(NO3)]N.7 Even though it is still in its infancy,
this spectroscopy appears to be a very sensitive probe of the
chiral molecular environment around the absorbing atom. An
important aspect of XNCD originates from a not well-known
electric dipole-electric quadrupole mechanism that exists only
in oriented samples. Moreover, it can only be observed in
crystals belonging to one of 13 noncentrosymmetric crystal
point groups.8 Unfortunately, the electric dipole-magnetic
dipole mechanism that is dominant in UV/vis natural circular
dichroism, and which is observable in isotropic chiral media,
vanishes in core-level X-ray spectroscopy. From this perspective,
a related phenomenon, X-ray magnetochiral dichroism (XMcD),
i.e. the propensity of chiral paramagnetic systems to exhibit
a differential absorption of unpolarized light in the presence of
a magnetic eld is more appealing. This peculiar type of
dichroism is due to orbital toroidal currents, which are relevant
to many phenomena, ranging from multiferroicity to superconductivity. In the X-ray range, this effect has been unambiguously demonstrated in only two cases: in chromium oxide,9
and more recently in [Co(hfac)2NITPhOMe]N.7b In addition to
these dichroic effects, there exists also X-ray natural linear
dichroism (XNLD), which directly measures the anisotropy of
the electronic states of an absorbing atom.
Due to the attractive optical3b and magnetic properties1 of
EMACs, a general method for their enantiomeric resolution is
highly desirable. In the present investigation, we have chosen to
study tricobalt complexes supported by 2,20 -dipyridylamine
ligands, not only due to their preparative ease and air-stability
in the solid state, but also for their interesting magnetic
This journal is © The Royal Society of Chemistry 2018
Chemical Science
properties. Indeed, the electronic structure of tricobalt EMACs
has been extensively studied, primarily due to the observation
that several [Co3(dpa)4X2] derivatives display a temperatureinduced S ¼ 1/2 to 3/2 spin-crossover (SCO).1c,10 DFT calculations reveal that, in the ground state, one unpaired electron is
delocalized over three equidistant cobalt centers in a metalbased snb molecular orbital.11 It is thought that SCO can then
proceed by two different mechanisms, characterized by the
retention or loss of the symmetry in the {Co3} core.1,12
Our synthetic approach takes advantage of the dicationic
character of [Co3(dpa)4(MeCN)2]2+,2 which permitted the selective precipitation of each of the enantiomers in association with
an enantiopure chiral anion. As this method involves
a straightforward anion exchange reaction, it is not necessary to
rely on serendipity or to synthesize chiral ligands. For this work,
we have chosen a pair of chiral arsenyl dianions, based on
inexpensive and easily available tartaric acid.13 Such D2symmetric As(III), as well as Sb(III), tartrate adducts have been
widely used in chiral HPLC separations14 and enantiopure
crystallization15 of metal complexes, and different mechanistic
explanations for the chiral recognition of octahedral metal
complexes have been proposed.16 It is worth mentioning that
potassium antimonyl tartrate is commercially available, but its
low solubility in organic solvents precluded its use.
We present here results on the use of the lipophilic arsenyl
tartrate salt, (NBu4)2[As2(tartrate)2] (D-1 and L-1) in the resolution of [Co3(dpa)4(MeCN)2]2+ enantiomers. Chiral compounds
so-obtained were optically, structurally and magnetically characterized and were found to display a space group and
a magnetic ground state compatible with the demonstration of
X-ray natural and magnetochiral optical activity. Results of
thorough polarization dependent X-ray spectroscopic studies
performed at the Co K-edge reveal large XNCD and XNLD
signals, but only vanishingly small XMCD and XMcD signals,
likely due to the strongly delocalized nature of the magnetization on the tricobalt core.
Results and discussion
Synthesis and characterization
The tricobalt EMAC salt [Co3(dpa)4(MeCN)2](BF4)2 was formed
in situ by stirring an acetonitrile solution of [Co3(dpa)4Cl2]17 and
2 eq. of AgBF4 for several hours. Aer ltration to remove AgCl,
1 eq. of D-1 or L-1 dissolved in MeCN was added to the ltrate
and stirred for 1 h, yielding a deep green precipitate. The solid
was isolated by ltration and re-dissolved in MeCN. Vapor ether
diffusion into the acetonitrile solutions gave green plates
(fast crystallization) and/or blocks (slow crystallization) of
[D-Co3(dpa)4(MeCN)2](NBu4)2[L-As2(tartrate)2]2$solvent, (D-2),
or
[L-Co3(dpa)4(MeCN)2](NBu4)2[D-As2(tartrate)2]2$solvent,
(L-2), aer 2 days in ca. 40% yield. This yield is quite satisfactory, considering a maximum yield of 50%, given that the
starting [Co3(dpa)4(MeCN)2](BF4)2 is racemic and the enantiomers are not expected to interconvert in acetonitrile (vide infra).2
[rac-Co3(dpa)4(MeCN)2][As2(tartrate)2]$2MeCN (rac-3) was also
obtained by a similar reaction, using a racemic mixture of
Chem. Sci., 2018, 9, 1136–1143 | 1137
View Article Online
Open Access Article. Published on 04 December 2017. Downloaded on 8/6/2019 6:35:12 PM.
This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.
Chemical Science
Edge Article
(NBu4)2[As2(tartrate)2] 1, and crystallized directly from acetonitrile upon standing in 95% yield.
The stoichiometry of 2 is unexpected, the simplest formula
being the binary salt [Co3(dpa)4(MeCN)2][As2(tartrate)2].
Instead, the compounds consist of one cobalt complex, two
tetrabutylammonium cations and two arsenate dianions. The
unusual stoichiometry may be explained by the 2-fold excess of
1 used in this anion exchange protocol. However, when 0.5 eq.
of D or L-(NBu4)2[As2(tartrate)2] was used, the resulting green
precipitate could not be subsequently crystallized and powder
diffraction indicated that the polycrystalline precipitate was not
isostructural with 2 (Fig. S11†).
The crystal structures of D-2 and L-2 were successfully solved
and rened in the non-centrosymmetric P4212 space group,
with Flack parameters of 0.008(2) and 0.002(2), respectively
(Table 1). The [Co3(dpa)4(MeCN)2]2+ moieties are parallel to one
another and are arranged so that the metal axes lie along the c
crystal direction (Fig. 1). The arsenyl tartrate anions nestle in
the voids between the molecular layers, with distances of 3.22
and 3.11 Å between the CO group of the arsenyl tartrate and
the axial MeCN and one pyridyl group of the EMAC, respectively
(Fig. S4†). Notably, [D-As2(tartrate)2]2 crystallizes exclusively
with [L-Co3(dpa)4(MeCN)2]2+ and vice versa; here, the D and L
notation refers to the handedness of the tricobalt helix. The Co–
Co distances are identical, due to a 2-fold axis bisecting the
central cobalt ion. Bond distances and angles are typical of the
starting materials and show no unusual features (Tables S1
and S2†).
The electron density of the interstitial solvent molecules was
highly diffuse and could not be satisfactorily modeled. Renement of the structures without solvent molecules gave values of
R1 (all data) ¼ 0.0553 and 0.0591 for D-2 and L-2, respectively.
The Platon Squeeze procedure18 calculated one void space per
unit cell (Fig. S7†), accounting for 39% of the total unit cell
volume and containing about 550 electrons. The electron count
corresponds to approximately twelve diethylether molecules per
Table 1
Arrangement of 2 in the crystal viewed along the crystalline
a axis; Co: dark blue, As: violet, O: red, N: light blue, C: gray, H: pink.
Fig. 1
unit cell, or six ether molecules per formula unit (Z ¼ 2).
Successful re-renement of the structure yielded R1 (all data) ¼
0.0223 and 0.0353 for D-2 and L-2, respectively. Thermogravimetric analysis (TGA) on a freshly ltered sample (Fig. S12†)
showed a mass loss of approximately 11%, corresponding to ca.
four molecules of diethylether per formula unit. Mass loss
begins even at room temperature, accounting for the discrepancy between the number of solvent molecules estimated by
X-ray diffraction and TGA. Solvent loss was also observed in the
loss of crystallinity during attempted data collection at 280 K. As
the void space is continuous in the ab plane, the solvent is not
only highly mobile within the crystal, but can easily escape the
crystal as well.
Compound rac-3 crystallizes in the centrosymmetric C2/c
space group and contains both enantiomers of the cations and
anions. The D and L forms are segregated and contrary to the
heterochiral interactions seen in D-2 and L-2, anions and
cations of the same chirality associate with each other in 3
Crystallographic data for D-2, L-2 and rac-3
T/K
Empirical formulaa
fw
Space group
a/Å
b/Å
c/Å
b/
V/Å3
Z
d (calc., g cm 3)
m (mm 1)
R indices (all data)
GooF on F2
Flack parameter
D-2
L-2
rac-3
120
C92H118As4Co3N16O24
2308.49
P4212 (#90)
20.6481(10)
20.6481(10)
16.8896(9)
90
7200.8(8)
2
1.065
1.309
R1b ¼ 0.0223, wR2c ¼ 0.0512
1.070
0.008(2)
120
C92H118As4Co3N16O24
2308.49
P4212 (#90)
20.6668(10)
20.6668(10)
16.8568(9)
90
7199.8(8)
2
1.065
1.309
R1b ¼ 0.0353, wR2c ¼ 0.0784
1.052
0.002(2)
120
C56H48As2Co3N16O12
1463.73
C2/c (#15)
17.794(4)
19.121(4)
19.042(4)
116.941(15)
5776(2)
4
1.683
2.067
R1b ¼ 0.0324, wR2c ¼ 0.0693
1.061
—
a
Solvent molecules of crystallization are not included in the empirical formula. b R1 ¼ S||Fo|
¼ 1/s2(Fo2) + (aP)2 + bP, where P ¼ [max(0 or Fo2) + 2(Fc2)]/3.
1138 | Chem. Sci., 2018, 9, 1136–1143
|Fc||/S|Fo|. c wR2 ¼ [S[w(Fo2
Fc2)2]/S[w(Fo2)2]]1/2, w
This journal is © The Royal Society of Chemistry 2018
View Article Online
Open Access Article. Published on 04 December 2017. Downloaded on 8/6/2019 6:35:12 PM.
This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.
Edge Article
(Fig. S5†), with distances of 3.14 and 3.12 Å between an Ascoordinated O atom of the arsenyl tartrate and an axial MeCN
and one pyridyl group of the EMAC, respectively (Fig. S6†).
The chirality of D-2 and L-2 was rst conrmed by UV/vis
circular dichroism spectroscopy in acetonitrile solution
(Fig. 2). Between 250 and 650 nm, the spectra plotted as D3 vs. l
are similar, but ca. 4 more intense, to those reported by Cotton
and coworkers for the spontaneously resolved [Co3(dpa)4(MeCN)2](PF6)2 enantiomers,2 although the 3 vs. l values in the
absorption spectra were found to be similar. The arsenyl tartrate
peaks below 250 mn are masked by the much more intense
peaks from the [Co3(dpa)4(MeCN)2]2+ moiety in the same area.19
Given that the circular dichroism spectra were measured
separately, yet exhibit the same amplitude with opposite sign,
fast racemization in solution can be excluded. A CD spectrum
on a MeCN solution of D-2 was obtained aer preparation, and
again aer 8 days at r.t. No loss of peak intensity was observed
(Fig. S17†). Finally, the potential for chemical manipulation of 2
without loss of enantiopurity was demonstrated by exchanging
the arsenyl tartrate anions with PF6 anions on a sample of D-2.
Recrystallization from acetonitrile and diethyl ether yielded
crystals of [D-Co3(dpa)4(MeCN)2](PF6)2$MeCN$Et2O, as previously obtained by Cotton et al. by spontaneous resolution (Table
S4, Fig. S8 and S9†).2 Circular dichroism spectroscopy in MeCN
of [D-Co3(dpa)4(MeCN)2](PF6)2 revealed slightly less intense
peaks compared to those of D-2, and no evolution aer 5 days
(Fig. S18†), demonstrating that the chirality of the compound is
maintained in solution, even in the absence of chiral anions.
To investigate potential SCO properties in D-2 and L-2, the
magnetic susceptibility was evaluated between 1.85 and 320 K
in a 0.1 T applied eld (Fig. 3 and S13†). Both compounds show
a constant cT product at about 0.50 cm3 K mol 1 between
5 and 200 K, close to what is expected for an S ¼ 1/2 system with
g ¼ 2.4 (0.54 cm3 K mol 1), typical of other trinuclear cobalt
EMACs.1c,2 Above 200 K, the cT product increases gradually, as
the signature of a SCO process, which becomes irreversible
above 320 K as a sign of decomposition, involving most likely
the loss of the interstitial solvent molecules. Low temperature
magnetization curves plotted vs. HT 1, were superimposable,
and a t of these data at 1.9 K to the S ¼ 1/2 Brillouin function,
Chemical Science
Temperature dependence of the cT product for D-2 (red) and
L-2 (blue) at 0.1 T, where c is the magnetic susceptibility defined as M/
H per mole of complex, with M being the magnetization and H the
applied magnetic field.
Fig. 3
gave g ¼ 2.3(1) for D-2 and g ¼ 2.2(1) for L-2 (Fig. S13†). The
uncertainty in these two values is due to the inevitable doubt
regarding the molecular weight of the samples because of
interstitial solvent loss. Altogether, these results suggest that
the delocalized 2Ag state proposed by theory is the main
contribution to electronic structure of 2 at all relevant
temperatures.
X-ray absorption
To investigate the local chiral and magnetic properties of D-2
and L-2 we have performed thorough polarization dependent Xray spectroscopy studies at the Co K-edge on single crystals. The
experiments were carried out at the ID12 beamline of the
European Synchrotron Radiation Facility, which is dedicated to
such studies in the hard X-ray range. As D-2 and L-2 are chiral
paramagnets, four different types of dichroisms in X-ray
absorption are allowed by symmetry: X-ray Natural Linear
Dichroism (XNLD) and X-ray Natural Circular Dichroism
(XNCD), while application of an external magnetic eld also
permits X-ray Magnetic Circular Dichroism (XMCD), and X-ray
Magneto-Chiral Dichroism (XMcD). If sH+,L represents the
cross-section measured with le circularly polarized X-rays (L)
in an external magnetic eld parallel (H+) to the X-ray propagation vector (and similar obvious denitions for right circular
polarization and an antiparallel magnetic eld), the three
dichroic cross-sections making use of circular polarization are
dened by
sXMCD ðHÞ ¼
sXMcD ðHÞ ¼
sXNCD ðHÞ ¼
UV-visible circular dichroism spectra of D-2 (red) and L-2
(blue) between 200–650 nm in acetonitrile solution at room
temperature.
Fig. 2
This journal is © The Royal Society of Chemistry 2018
½sH þ ;R þ sH
;L
sH
;R
sH þ ;L
sH
;R
sH
;L
sH þ ;L
sH
;L
2
½sH þ ;R þ sH þ ;L
2
½sH þ ;R þ sH
;R
2
(1)
(2)
(3)
It should be mentioned that sXNCD(H), being independent of
the magnetic eld, is usually measured for H ¼ 0. These three
dichroic cross-sections are independent from each other and
Chem. Sci., 2018, 9, 1136–1143 | 1139
View Article Online
Open Access Article. Published on 04 December 2017. Downloaded on 8/6/2019 6:35:12 PM.
This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.
Chemical Science
Edge Article
provide information on the electronic and magnetic structures
of the element-specic absorbing atoms.
Our rst experiment was an attempt to measure sXMCD. A
single crystal of D-2 was cooled down to 3 K and placed in the
bore of a superconducting solenoid providing a magnetic eld
up to 17 T. Absorption spectra were collected using circularly
polarized X-rays where the propagation vector of the X-rays, as
well as the magnetic eld, were aligned along the c axis of the D2 crystal. In the present experiment, the XMCD signal was too
weak to be detected, even under an applied magnetic eld of 17
T (Fig. S14†). From the noise level in the XMCD spectra, we
could only give an estimated upper limit of the signal, <0.2%,
with respect to the edge jump. Whenever a crystal exhibits X-ray
natural circular dichroism, the true XMCD signal can be obtained only as a combination of four X-ray absorption spectra
measured independently (eqn (1)). In our case, the XMCD signal
can be seen as the difference between two XNCD signals
measured with opposite directions of applied magnetic eld,
sXMCD(H) ¼ sXNCD(H+) sXNCD(H ). Since the crystals are chiral,
both sXNCD(H+) and sXNCD(H ) are “large” with similar intensities, so one indeed expects the XMCD signal to be difficult to
detect. On the other hand, the magnetic moment of D-2 and L-2
originating from the three metal–metal bonded cobalt atoms is
quite small. Out of the 21 3d electrons, all are paired, except for
one that lies in a highly delocalized SOMO with snb symmetry.11
With three cobalt atoms and a maximum magnetic moment of
1 mB, and no 3d density on the central cobalt atom due to
symmetry reasons, an individual cobalt atom carries
a maximum moment of 0.5 mB. In addition to the small average
magnetic moment, DFT calculations on the parent dichloride
derivative showed that the SOMO presents substantial density
on the chlorine atoms, reducing even further the average
magnetic polarization of the cobalt atoms.12 This can be
compared to more common situations where high-spin divalent
cobalt atoms carry a spin magnetic moment of 3 mB with some
additional orbital magnetic moment.7b,20 When comparing with
previous Co K-edge XMCD data from literature, e.g. Co2+ ions in
a ZnO crystal, a simple cross-multiplication suggests that the
expected intensity for XMCD in sample 2 cannot be larger than 2
10 4 of the XAS spectra.20a This small signal would already be
quite difficult to measure in non-chiral crystals, but in the
present case it is completely obscured by the presence of the
XNCD signals that, as we will show, are as large as 2% of the XAS
cross-sections.
In order to check for the absence of structural distortions in
these crystals, we further performed X-ray Natural Linear
Dichroism (XNLD) measurements (Fig. 4). In non-cubic crystals
possessing a symmetry axis of an order larger or equal to three,
the XNLD signal is measured by
sXNLD ¼ st
in the orthoaxial conguration, with the c axis perpendicular to
the propagation vector of the X-rays (the q ¼ 90 geometry in
Scheme 1), and recorded the spectra for two orthogonal linearly
polarized X-rays, 3p and 3s. In this experimental geometry, 3p is
parallel to the 4-fold axis, whereas 3s is perpendicular to it. A
large XNLD signal, amounting to 70% of the XAS cross-section,
was obtained (Fig. 4b). To further conrm the 4-fold axis
symmetry, we mounted a crystal of L-2 in the axial conguration, with the 4-fold axis parallel to the X-ray propagation vector
(the q ¼ 0 geometry in Scheme 1), and again recorded spectra
with 3p and 3s linear polarizations. Any distortion reducing the
4-fold axis to 2-fold would be detected by the presence of
a difference signal in this geometry and Fig. 4a clearly shows the
absence of such signal thus conrming the 4-fold symmetry of
the crystal.
In order to verify that the crystals of D-2 and L-2 exhibit the
expected chirality, we performed XNCD for several crystals of
both enantiomers using circularly polarized X-rays (Fig. 5a).
Whenever the spectra are recorded under an applied magnetic
eld or the sample has a nite magnetization, the true XNCD
spectra should be calculated as a half-sum of the difference of the
absorption spectra obtained with le and right circular polarizations for two directions of the applied magnetic eld (see eqn
(3)). Since the XNCD is independent of magnetic eld and our
samples are not magnetically ordered, the true XNCD signal is
given by sL–sR. Here the experiments were performed without an
sk
where st and sk are the cross-sections measured with the linear
polarization vector 3 perpendicular or parallel, respectively, to
the symmetry axis. For such crystals, the main contribution
comes from the electric dipole term and one usually neglects
the electric quadrupole term. We rst mounted a crystal of L-2
1140 | Chem. Sci., 2018, 9, 1136–1143
XANES and XNLD spectra obtained with linearly polarized
X-rays for (a) axial (q ¼ 0 ) and (b) orthoaxial (q ¼ 90 ) configurations.
Fig. 4
Scheme 1 Schematic diagram of the measurement geometry
showing the relationship between the tetragonal crystal axes and the
direction and polarization of the X-ray beams. Axial geometry corresponds to the X-ray wavevector being parallel (q ¼ 0 ) to the crystal c
axis whereas in the orthoaxial configuration the X-ray propagation
direction is perpendicular to the optical axis of the crystal.
This journal is © The Royal Society of Chemistry 2018
View Article Online
Open Access Article. Published on 04 December 2017. Downloaded on 8/6/2019 6:35:12 PM.
This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.
Edge Article
(a) XANES (black) and XNCD spectra obtained in axial configuration (q ¼ 0 ) for D-2 (red) and L-2 (blue); (b) comparison of axial (blue)
and orthoaxial (green) XNCD spectra for an oriented crystal of L-2.
Fig. 5
Chemical Science
In the XMCD experiments, all four spectra sH+,L, sH ,L, sH+,R
and sH ,R have been recorded, and it is thus also possible to
extract the X-ray magneto-chiral dichroism. It was however
found that sXMcD cannot be distinguished from the background
noise (Fig. S14†). For XMcD to be present at the K-edge, one
needs both a large orbital magnetic polarization of the
absorbing atom and a strong p–d orbital hybridization. The p–
d hybridization is present, as demonstrated by XNCD signal
that is larger than 1% between the pre-edge feature up to 25 eV
above the main rising edge. On the other hand, the rather small
magnetic polarization that already precluded us from
measuring a sizeable XMCD effect is very likely responsible for
the absence of detectable magnetochiral effect.
Conclusions
applied magnetic eld and the XNCD spectra were calculated as
the difference of the absorption spectra obtained with le and
right circular polarizations. As shown in Fig. 5a, with the crystal c
axis mounted parallel to the incident X-ray beam (q ¼ 0 ), the
cobalt enantiomers in D-2 and L-2 display the expected mutual
mirror-image symmetry. The axial XNCD signals are rather large,
with intensities up to about 2% of the absorption cross-section.
Given that the intensity of the XNCD signals are equivalent for
both enantiomers over several measured samples, we can
conclude that the crystals are effectively enantiopure.
X-ray natural circular dichroism stems from the interference
terms which mix multipole transitions of opposite parity: the
electric dipole–electric quadrupole, i.e. 1 s / (p,d) in the case of
the K-edge. As one can see from Fig. 5, the XNCD spectrum in 2
covers a wide energy range (more than 50 eV above the
absorption edge), indicating that spatially extended d orbitals in
the continuum hybridized with p-states of the Co atom also
contribute, together with more localized 3d–4p orbitals.
As opposed to its analogue in the visible range, XNCD has
a more complicated angular dependence and is observable only
in 13 non-centrosymmetric crystal point groups. For the point
group D4, symmetry requires that the angular dependence of
XNCD varies as a function of a single spectral shape, sXNCD
multiplied by the angular factor (3 cos2 q 1), where q is the
polar angle between the crystal c axis, the optical axis, and the
wave vector k⃑.21 This dependence can be indeed observed in the
XNCD spectra when varying the q angle (Fig. S15†), and is best
observed by plotting the XNCD signals for axial q ¼ 0 and
orthoaxial q ¼ 90 congurations (Fig. 5b) where one observes
that sXNCD (q ¼ 90 ) ¼ 1/2 sXNCD (q ¼ 0 ), as expected for the
angular dependence of (3 cos2 q 1) predicted by group theory.
Circularly polarized light is the 90 phase-shied sum of
vertical and perpendicular polarizations. When q ¼ 90 , half of
the circularly polarized X-rays is parallel to the crystalline c axis,
and half is perpendicular. Accordingly, the XANES spectrum
obtained with circularly polarized light is equivalent to the
averaged spectra obtained from horizontal and vertical polarization (Fig. S16†). Conversely, when q ¼ 0 , the circular polarization of X-rays is always perpendicular to the c axis and is thus
equivalent to the case with q ¼ 90 under horizontal linear
polarization (Fig. S16†).
This journal is © The Royal Society of Chemistry 2018
We have used the chiral anion [As2(tartrate)2]2 to easily resolve
racemic mixtures of [Co3(dpa)4]2+ cations by selective crystallization. Both EMAC enantiomers crystallize as [Co3(dpa)4(MeCN)2](NBu4)2[As2(tartrate)2]2 in the non-centrosymmetric
space group P4212 with Flack parameters statistically equal to
zero. The chiral EMACs display persistent optical activity in
solution, as evidenced by mirror-image circular dichroism
spectra in MeCN. Single crystals of [Co3(dpa)4(MeCN)2](NBu4)2[As2(tartrate)2]2 furthermore display substantial optical
activity in the hard X-ray range, giving mirror-image X-ray
natural circular dichroism at the cobalt K-edge. The angular
dependence of the XNCD and XNLD spectra is consistent with
what is expected for a system with D4 symmetry. Unfortunately,
the strongly delocalized nature of the magnetism in these
systems did not allow us to record XMCD and XMcD signals
even under an applied magnetic eld of 17 T. Nonetheless,
these results demonstrate that polarized X-rays provide an
original approach for the investigation of chirality and
symmetry in single crystals.
We are now examining the efficacy of our synthetic method
for the separation of chiral EMACs with higher nuclearity and
different metal centers (e.g. chromium, nickel and copper).
With potentially higher specic rotation with increasing
nuclearity, ground states with a larger magnetic moment, and
catalytically active axial sites, chiral EMACs open interesting
perspectives in non-linear optics, magneto-optical effects and
asymmetric synthesis.
Experimental section
Materials and methods
All manipulations were carried out under dry nitrogen or argon
using standard drybox or Schlenk techniques. Tetrahydrofuran,
dichloromethane, diethyl ether and acetonitrile were puried
using an Innovative Technologies solvent purication system.
CoCl2 was purchased from Fisher Chemicals and dried for
several days at 130 C before use. 2,20 -Dipyridylamine was
purchased from Aldrich and recrystallized in hexane. AgBF4 was
purchased from Strem Chemicals and stored in a nitrogen
glovebox. [Co3(dpa)4Cl2],15 D-1 and L-119 were prepared as reported elsewhere.
Chem. Sci., 2018, 9, 1136–1143 | 1141
View Article Online
Chemical Science
Open Access Article. Published on 04 December 2017. Downloaded on 8/6/2019 6:35:12 PM.
This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.
Synthesis
General synthesis for 2. [Co3(dpa)4Cl2] (0.15 g, 0.16 mmol)
and AgBF4 (0.07 g, 0.36 mmol) in 2 mL of MeCN were stirred at
room temperature for 16 h. The resulting deep green solution
was ltered using VWR plastic lter (0.2 mm) and a solution of
D-1 or L-1 (0.16 g, 0.17 mmol) in 1 mL of MeCN was added. The
solution was stirred for 1 h and the resulting solid was ltered
off, dissolved in MeCN and recrystallized by ether vapor
diffusion.
[D-Co3(dpa)4(MeCN)2](NBu4)2[L-As2(tartrate)2]2$solvent, D2. Yield: 42% (0.16 g). Elemental analysis calc. for C92H118As4Co3N16O24(C4H10O)0.5(H2O) (2363.61 g mol 1) C 47.77, H 5.33, N
9.48; found C 47.31, H 5.55, N 9.28. IR (ATR, cm 1): 2273 w,
1657 s, 1606 s, 1592 s, 1548 w, 1471 s, 1457 s, 1426 s, 1371 m,
1314 m, 1283 w, 1265 w, 1155 s, 1125 s, 1071 m, 1023 s, 899 s,
840 m, 808 m, 761 m, 733 s, 661 m, 573 m.
[L-Co3(dpa)4(MeCN)2](NBu4)2[D-As2(tartrate)2]2$solvent, L2. Yield: 41% (0.15 g). Elemental analysis calc. for C92H118As4Co3N16O24(C4H10O)0.5 (2345.59 g mol 1) C 48.13, H 5.29, N 9.55;
found C 48.18, H 5.53, N 9.22. IR (ATR, cm 1): 2280 w, 1606 s,
1593 s, 1549 w, 1470 s, 1457 s, 1426 s, 1371 m, 1314 m, 1282 w,
1264 w, 1155 s, 1125 s, 1071 m, 1023 s, 900 s, 841 m, 807 m, 761
m, 732 s, 660 m, 573 m.
[Co3(dpa)4(MeCN)2][As2(tartrate)2]$2MeCN, rac-3. [Co3(dpa)4Cl2] (0.05 g, 0.05 mmol), AgBF4 (0.02 g, 0.1 mmol) and
10 mL of MeCN were stirred at room temperature for 16 h to give
a deep green solution and a white precipitate. The resulting
solution was ltered using VWR plastic lter (0.2 mm) and
a mixture of D-1 (0.03 g, 0.03 mmol) and L-1 (0.03 g, 0.03 mmol)
was added. Overnight, large brown blocks were obtained. Yield
95% (0.07 g). Elemental analysis calc. for C52H42As2Co3N14O12(MeCN)0.25(H2O)1.35 (1416.22 g mol 1) C 44.53, H 3.23, N
14.09; found C 44.79, H 3.50, N 13.85. IR (ATR, cm 1): 2264 w,
1659 s, 1604 s, 1594 s, 1548 w, 1454 s, 1422 s, 1372 m, 1335 m,
1314 m, 1269 w, 1169 w, 1157 w, 1150 w, 1128 m, 1073 m, 1021
m, 920 w, 898 w, 885 w, 850 w, 840 w 806 m, 767 w, 757 m, 732 s,
659 w, 633 m, 572 w.
Physical measurements
CHNS elemental analyses were performed by the Service d'Analyse
Elementaire in Nancy. IR spectra were measured on a Nicolet 6700
FT-IR equipped with a SMART iTR™ accessory. Magnetic
susceptibility measurements were obtained using a Quantum
Design MPMS-XL SQUID magnetometer. The measurements were
performed on a polycrystalline sample of D-2 (35.9 mg) or L-2
(16.7 mg) introduced in a polyethylene bag (3 0.5 0.02 cm).
The dc measurements were conducted from 400 to 1.8 K at 0.1 and
1 T applied dc eld. An M vs. H measurement was performed at
100 K to conrm the absence of ferromagnetic impurities. CD
spectra in solution were obtained on a Jasco J-815 spectrophotometer. The measurements were made on samples with a 1.72
10 5 M and 1.62 10 5 M concentration for D-2 and L-2,
respectively. These samples were prepared by dissolving either
0.992 mg of D-2 or 0.936 mg of L-2 in 25 mL of MeCN. Class A
volumetric asks and a Mettler MX5 microbalance with an estimated weighting error of 2 mg were employed in the sample
1142 | Chem. Sci., 2018, 9, 1136–1143
Edge Article
preparation. All spectra were measured with a scan speed of 50
nm min 1 and a 2 nm bandwidth. Data in millidegree were converted to the differential extinction coefficient (D3) using the
equation:
D3 ¼
q
32980 c l
where q is the measured CD ellipticity in millidegrees, c is the
concentration in mol L 1, and l is the pathlength of the cuvette
in cm.
Crystallography
Crystals suitable for X-ray diffraction were selected under
immersion oil in ambient conditions and attached to a MiTeGen
microloop. The crystals were mounted in a stream of cold
nitrogen at 120(2) K and centered in the X-ray beam using a video
camera. Data collections were performed with Mo Ka (l ¼
0.71073 Å) radiation on a Bruker Quazar SMART APEX II, operating at 50 kV and 30 mA using graphite monochromated radiation. The data were collected using a routine to survey reciprocal
space, and reduction was performed using soware included in
Bruker Apex2 suite.22 The structures were solved using direct
methods.23,24 Non-hydrogen atoms were rened anisotropically
using weighted full-matrix least-squares on F2, followed by
difference Fourier synthesis.25,26 All hydrogen atoms were
included in the nal structure factor calculation at idealized
positions and were allowed to ride on the neighboring atoms with
relative isotropic displacement coefficients. The Platon SQUEEZE
routine18 was used to account for diffuse electron density arising
from interstitial solvent, which could not be modeled atomistically. Powder X-ray diffraction data were recorded using a PANalytical X'Pert PRO MPD diffractometer with Bragg–Brentano
geometry, Cu-Ka radiation (l ¼ 1.54184 Å) and a graphite back
scattering monochromator, using glass capillaries.
X-ray absorption
X-ray absorption experiments at the Co K-edge were carried out
at the ID12 beamline of the European Synchrotron Radiation
Facility (Grenoble, France), dedicated to polarization dependent
X-ray spectroscopy in the 2–15 keV energy range. A high ux of
circularly polarized photons was provided by an HELIOS-II
undulator. X-rays were monochromatized by a Si h111i double
crystal monochromator. Linearly polarized X-rays were obtained
using a 0.9 mm thick diamond quarter wave plate inserted
downstream with respect to the monochromator. For XNCD and
XNLD studies, the samples were mounted on an aluminum
sample holder and oriented so that the crystal's c axis and the
photons wave vector formed an angle of 0 or 90 . All spectra
were recorded at room temperature in total X-ray uorescence
yield detection mode in backscattering geometry. XNLD spectra
were obtained by ipping the two orthogonal linear polarizations of the X-rays at every point of the energy scan. Each set of
XNCD spectra were acquired by alternating the photon helicity
aer each spectrum 6 times, for a total of 12 absorption spectra.
XMCD and XMcD spectra were obtained by also using the
total uorescence yield detection mode in backscattering
This journal is © The Royal Society of Chemistry 2018
View Article Online
Open Access Article. Published on 04 December 2017. Downloaded on 8/6/2019 6:35:12 PM.
This article is licensed under a Creative Commons Attribution-NonCommercial 3.0 Unported Licence.
Edge Article
geometry, but under an applied magnetic eld. The samples
were mounted on the cold nger of a constant-ow helium
cryostat inserted in the bore of a superconducting solenoid
producing a magnetic eld up to 17 T. The temperature of the
samples were kept at 3.1 K. Under these experimental conditions, the samples were magnetically saturated.
Conflicts of interest
There are no conicts to declare.
Acknowledgements
This work was supported by the CNRS, the University of
Bordeaux, the Conseil Régional de la Nouvelle Aquitaine, the
France Canada Research fund (PhD fellowship for AS), the
European Union's Horizon 2020 research and innovation
program under the Marie Sklodowska-Curie grant agreement
No 706556 CHIMMM (Postdoctoral fellowship for MC), the
Erasmus Mundus European program (PhD fellowship for VB),
the ANR project CHIROTS ANR-11-JS07-013-01 (PR, PhD
fellowship for AN), the European Synchrotron Radiation Facility
(beamtimes CH-4171, CH-4457 and CH-5009), the GDR MCM-2
and the MOLSPIN COST action CA15128. The authors warmly
thank K. Ollefs, F. Guillou, A. Hen, P. Voisin, M. Rouzières, L.
Etienne and D. Denux for technical assistance and L. Falvello
for useful discussions.
References
1 (a) S.-A. Hua, M.-C. Cheng, C.-h. Chen and S.-M. Peng, Eur. J.
Inorg. Chem., 2015, 2510; (b) M. Majumdar and J. K. Bera,
Supramolecular and Self-Assembled Metal-Containing
Materials, in Macromolecules Containing Metal and MetalLike Elements, ed. A. Abd-El Aziz, C. E. Carraher, C. U.
Pittman and M.Zeldin, John Wiley & Sons, Inc., 2009, vol.
9, pp. 182–253; (c) J. F. Berry, in Multiple Bonds between
Metal Atoms, ed. F. A. Cotton, C. A. Murillo and R. A.
Walton, Springer U.S., 2005, pp. 669–706.
2 R. Clérac, F. A. Cotton, K. R. Dunbar, T. Lu, C. A. Murillo and
X. Wang, Inorg. Chem., 2000, 39, 3065.
3 (a) D. W. Armstrong, F. A. Cotton, A. G. Petrovic,
P. L. Polavarapu and M. M. Warnke, Inorg. Chem., 2007, 46,
1535; (b) M. M. Warnke, F. A. Cotton and D. W. Armstrong,
Chirality, 2007, 19, 179.
4 C.-H. Yu, M.-S. Kuo, C.-Y. Chuang, G.-H. Lee, S.-A. Hua,
B.-Y. Jin and S.-M. Peng, Chem.–Asian J., 2014, 9, 3111.
5 R. A. Pascal Jr and A. P. West Jr, Tetrahedron, 2013, 69, 6108.
6 (a) J. Goulon, C. Goulon-Ginet, A. Rogalev, V. Gotte,
C. Malgrange, C. Brouder and C. R. Natoli, J. Chem. Phys.,
1998, 108, 6394; (b) L. Alagna, T. Prosperi, S. Turchini,
J. Goulon, A. Rogalev, C. Goulon-Ginet, C. R. Natoli,
R. D. Peacock and B. Stewart, Phys. Rev. Lett., 1998, 80, 4799.
7 (a) B. Stewart, R. D. Peacock, L. Alagna, T. Prosperi,
S. Turchini, J. Goulon, A. Rogalev and C. Goulon-Ginet, J.
Am. Chem. Soc., 1999, 121, 10233; (b) R. Sessoli,
M.-E. Boulon, A. Caneschi, M. Mannini, L. Poggini,
This journal is © The Royal Society of Chemistry 2018
Chemical Science
F. Wilhelm and A. Rogalev, Nat. Phys., 2015, 11, 69; (c)
I. Mihalcea, M. Perfetti, F. Pineider, L. Tesi, V. Mereacre,
F. Wilhelm, A. Rogalev, C. E. Anson, A. K. Powell and
R. Sessoli, Inorg. Chem., 2016, 55, 10068.
8 J. Goulon, C. Goulon-Ginet, A. Rogalev, G. Benayoun,
C. Brouder and C. R. Natoli, J. Synchrotron Radiat., 2000, 7, 182.
9 J. Goulon, A. Rogalev, F. Wilhelm, C. Goulon-Ginet, P. Carra,
D. Cabaret and C. Brouder, Phys. Rev. Lett., 2002, 88, 2374011.
10 R. Clérac, F. A. Cotton, L. M. Daniels, K. R. Dunbar,
K. Kirschbaum, C. A. Murillo, A. A. Pinkerton, A. J. Schultz
and X. Wang, J. Am. Chem. Soc., 2000, 122, 6226.
11 M.-M. Rohmer and M. Bénard, J. Am. Chem. Soc., 1998, 120,
9372.
12 M.-M. Rohmer, A. Strich, M. Bénard and J.-P. Malrieu, J. Am.
Chem. Soc., 2001, 123, 9126.
13 D. Marcovich and R. E. Tapscott, J. Am. Chem. Soc., 1980, 102,
5712.
14 (a) S. Yamazaki and H. Yoneda, Inorg. Nucl. Chem. Lett., 1979,
15, 195; (b) H. Nakazawa and H. Yoneda, J. Chromatogr.,
1978, 160, 89; (c) J. E. Harris, N. Desai, K. E. Seaver,
R. T. Watson, N. A. Kane-Maguire and J. F. Wheeler, J.
Chromatogr. A, 2001, 919, 427.
15 (a) B. D. Sarma Jr and J. C. Bailar, J. Am. Chem. Soc., 1955, 77,
5480; (b) U. Sakaguchi, K. Tomioka, T. Kashihara and
H. Yoneda, Inorg. Chem., 1985, 24, 463; (c) J. D. Harper,
K. A. Kane-Maguire, L. A. P. Kane-Maguire and
N. A. P. Kane-Maguire, Inorg. Chim. Acta, 1996, 249, 5.
16 (a) A. Zalkin, D. H. Templeton and T. Ueki, Inorg. Chem., 1973,
12, 1641; (b) T. Bark, A. von Zelewsky, D. Rappoport,
M. Neuburger, S. Schaffner, J. Lacour and J. Jodry, Chem.–
Eur. J., 2004, 10, 4839; (c) K. Miyoshi, S. Izumoto, K. Nakai
and H. Yoneda, Inorg. Chem., 1986, 25, 4654; (d) K. Miyoshi,
S. Izumoto and H. Yoneda, Bull. Chem. Soc. Jpn., 1986, 59, 3475.
17 F. A. Cotton, L. M. Daniels and G. T. Jordan IV, Chem.
Commun., 1997, 421.
18 A. L. Spek, Acta Crystallogr., Sect. C: Struct. Chem., 2015, 71, 9.
19 A. Naim, Y. Bouhadja, M. Cortijo, A. Valentı́n-Pérez,
E. Duverger-Nédellec, H. D. Flack, E. Freysz, P. Guionneau,
A. Iazzolino, A. Ould Hamouda, P. Rosa, O. Stefańczyk, and
M. Zeggar, submitted.
20 (a) V. Ney, B. Henne, J. Lumetzberger, F. Wilhelm, K. Ollefs,
A. Rogalev, A. Kovacs, M. Kieschnick and A. Ney, Phys. Rev. B,
2016, 94, 224405; (b) B. Henne, V. Ney, J. Lumetzberger,
K. Ollefs, F. Wilhelm, A. Rogalev and A. Ney, Phys. Rev. B,
2017, 95, 05446.
21 C. R. Natoli, C. Brouder, P. Sainctavit, J. Goulon, C. GoulonGinet and A. Rogalev, Eur. Phys. J. B, 1998, 4, 1.
22 Bruker-AXS APEX2, SADABS, and SAINT Soware Reference
Manuals, Madison, Wisconsin, USA, 2009.
23 A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano,
C. Giacovazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori
and R. Spagna, J. Appl. Crystallogr., 1999, 32, 115.
24 G. M. Sheldrick, Acta Crystallogr., Sect. A: Found. Crystallogr.,
2008, 64, 112.
25 O. V. Dolomanov, L. J. Bourhis, R. J. Gildea, J. A. K. Howard
and H. Puschmann, J. Appl. Crystallogr., 2009, 42, 339.
26 L. J. Farrugia, J. Appl. Crystallogr., 2012, 45, 84.
Chem. Sci., 2018, 9, 1136–1143 | 1143